Constructing a borderland: Roman imperial geographic writers on Mesopotamia from the 1st to the 4th centuries CE

PDF

Download

Open document

Transcript (if available)

Content CONSTRUCTING A BORDERLAND ROMAN IMPERIAL GEOGRAPHIC WRITERS ON MESOPOTAMIA FROM THE 1 ST TO THE 4 TH CENTURIES CE Hamish Cameron 1/448 Table of Contents Acknowledgements...............................................................................................................................................6 Introduction...........................................................................................................................................................7 1. Borderlands...................................................................................................................................................8 2. Study Area..................................................................................................................................................16 “Mesopotamia” and the Mesopotamian Borderland................................................................................18 The First Four Centuries...........................................................................................................................20 Physical Geography..................................................................................................................................22 A Note on the Maps..................................................................................................................................22 3. Geography..................................................................................................................................................23 Geography in the Roman Empire............................................................................................................25 Roman Imperial Geographic Writers.......................................................................................................25 Scientific and Descriptive Geography......................................................................................................26 The Scope of Geography..........................................................................................................................27 Linear Conceptions of Space....................................................................................................................30 The Main Sources.....................................................................................................................................32 Strabo....................................................................................................................................................32 Pliny the Elder......................................................................................................................................33 Ptolemy.................................................................................................................................................34 Expositio...............................................................................................................................................34 Ammianus Marcellinus........................................................................................................................35 4. Discussing the Chapters.............................................................................................................................35 Part I: Constructing Space........................................................................................................................36 Chapter 1: Naming and Placing..........................................................................................................36 Chapter 2: States in the Borderland.....................................................................................................36 Part II: The Use of Knowledge.................................................................................................................37 Chapter 3: Hellenistic Knowledge.......................................................................................................37 Chapter 4: Roman Geography............................................................................................................37 Part III: The Mobile Landscape................................................................................................................38 Chapter 5: Commercial Factors...........................................................................................................38 Chapter 6: Conclusion.........................................................................................................................38 Chapter 1: Placing and Naming.........................................................................................................................40 Strabo..............................................................................................................................................................41 Οἱ Ἀσσύριοι.............................................................................................................................................41 Syria...........................................................................................................................................................48 Mesopotamia.............................................................................................................................................57 Mygdonia..................................................................................................................................................62 Gordyene....................................................................................................................................................68 Southern Parts............................................................................................................................................72 Conclusion.................................................................................................................................................78 Pliny the Elder................................................................................................................................................79 Book Five: Syria........................................................................................................................................80 Northwest Mesopotamia......................................................................................................................84 Commagene.........................................................................................................................................93 The Euphrates......................................................................................................................................95 2/448 Palmyra.................................................................................................................................................98 Book Six: Mesopotamia..........................................................................................................................101 Pliny’s Mesopotamian Lists................................................................................................................102 Adiabene and Parapotamia.................................................................................................................109 Conclusion...............................................................................................................................................112 Claudius Ptolemy.........................................................................................................................................113 Mesopotamia...........................................................................................................................................114 Adjacent Areas.........................................................................................................................................119 The Expositio Totius Mundi et Gentium....................................................................................................123 Ammianus Marcellinus................................................................................................................................127 The Eastern Provinces.............................................................................................................................128 The Persian Regiones..............................................................................................................................129 The Geography of Julian’s campaign (363-364 CE).............................................................................136 Conclusion...............................................................................................................................................141 Defining and Describing the Mesopotamian Borderland..........................................................................142 Chapter 2: States in the Borderland..................................................................................................................152 Network Empires.........................................................................................................................................152 The Problem.......................................................................................................................................152 States as Networks..............................................................................................................................154 Network Infrastructure of the Mesopotamian Borderland...............................................................158 Commagene in the Borderland..............................................................................................................160 Osrhoene in the Borderland....................................................................................................................171 Nisibis......................................................................................................................................................181 Palmyra as a borderland city...................................................................................................................190 Conclusions..................................................................................................................................................195 Chapter 3: Using Hellenistic Knowledge.........................................................................................................197 Strabo’s sources.............................................................................................................................................199 Pliny’s Sources..............................................................................................................................................213 Ptolemy’s Sources.........................................................................................................................................220 The Hellenistic Imprint on Mesopotamia...................................................................................................225 Spatial Measurement................................................................................................................................225 Delimitation and Denomination............................................................................................................230 Geographical Material.............................................................................................................................232 Conclusion....................................................................................................................................................237 Chapter 4: The Representation of Imperial Power...........................................................................................239 Strabo Drawing Borders...............................................................................................................................240 Semiramis................................................................................................................................................240 Alexander and the Macedonians.............................................................................................................247 Representing the Parthians......................................................................................................................252 Pliny (Dis)ordering Space............................................................................................................................261 Book Five: Syria......................................................................................................................................261 Book Six: Mesopotamia..........................................................................................................................265 Imperial Space...............................................................................................................................................268 Conclusion....................................................................................................................................................273 Chapter 5: Commercial Movement in the Borderland.....................................................................................275 The Productivity of Mesopotamia...............................................................................................................277 3/448 Long-distance trade......................................................................................................................................280 Strabo and the Euphrates Route..................................................................................................................284 Northern Routes...........................................................................................................................................297 Expositio Totius Mundi et Gentium............................................................................................................304 Desert Routes................................................................................................................................................308 North-South Routes.....................................................................................................................................320 Local Routes.................................................................................................................................................323 Skenitai..........................................................................................................................................................330 Controlling Mobile Populations.............................................................................................................334 Individual mobility..................................................................................................................................336 Representing Mesopotamian Trade.............................................................................................................337 Chapter 6: Conclusions: The Borderland in a Global Context........................................................................340 Mesopotamia as a Frontier............................................................................................................................349 Global Connections......................................................................................................................................352 Globalisation and Networks in the Mesopotamian Borderland.............................................................352 The Conceptualisation of the Mesopotamian Borderland.....................................................................355 Conclusion....................................................................................................................................................357 Appendix 1: A Gazetteer of Select Sites in the Mesopotamian Borderland.....................................................359 A1.1: The Eastern Bounds of Οἱ Ἀσσύριοι................................................................................................359 Chalonitis.................................................................................................................................................363 The Adiabene Section.............................................................................................................................367 A1.2: Pliny’s Syrian Lists..............................................................................................................................371 A1.3: Anthemusia/Batnae............................................................................................................................378 A1.4: Bambyke/Hierapolis...........................................................................................................................380 A1.5: Beroea.................................................................................................................................................381 A1.6: Carrhae...............................................................................................................................................381 A1.7: Chordiraza...........................................................................................................................................383 A1.8: Circesium............................................................................................................................................384 A1.9: Dura Europus......................................................................................................................................385 A1.10: Edessa................................................................................................................................................386 A1.11: Europus.............................................................................................................................................387 A1.12: Naarmalcha and Environs................................................................................................................389 A1.13: Nikephorion......................................................................................................................................391 A1.14: Nisibis...............................................................................................................................................392 A1.15: Palmyra.............................................................................................................................................393 A1.16: Singara..............................................................................................................................................394 A1.17: Sinnaca..............................................................................................................................................395 A1.18: Thapsacus..........................................................................................................................................403 A1.19: Tigranocerta.....................................................................................................................................407 The Sources...................................................................................................................................410 The Sites........................................................................................................................................415 Lucullus’ Campaign......................................................................................................................419 Reconciliation................................................................................................................................426 A1.20: Zeugma.............................................................................................................................................430 Bibliography......................................................................................................................................................431 4/448 Map Index Map 1: The Mesopotamian Borderland..............................................................................................................20 Map 2: Strabo's Assyria........................................................................................................................................42 Map 3: Strabo’s Syria...........................................................................................................................................50 Map 4: The Geographical Boundaries of Strabo's Mesopotamia......................................................................59 Map 5: Strabo's Mygdonia..................................................................................................................................64 Map 6: Gordyene.................................................................................................................................................70 Map 7: The 200mm Isohyet................................................................................................................................73 Map 8: The Skenitai and their Lands..................................................................................................................75 Map 9: Pliny's Syrian Overview.........................................................................................................................82 Map 10: Pliny's Syria...........................................................................................................................................83 Map 11: Mesopotamian Sites in Pliny NH 5.....................................................................................................87 Map 12: Pliny's Euphrates Itinerary....................................................................................................................97 Map 13: Desert Ports in Pliny...........................................................................................................................100 Map 14: Pliny's Mesopotamian Tribes..............................................................................................................103 Map 15: Pliny's Mountain Tribes......................................................................................................................105 Map 16: Mesopotamian Sites in Pliny NH 6...................................................................................................108 Map 17: Ptolemy's Mesopotamia......................................................................................................................117 Map 18: Ptolemy's Syrian Districts...................................................................................................................120 Map 19: Ammianus' Ciruitus............................................................................................................................131 Map 20: Adiabene in Context...........................................................................................................................134 Map 21: Known Sites from Julian's Campaign................................................................................................137 Map 22: Sites in the Mesopotamian Borderland..............................................................................................159 Map 23: Pliny's Syria.........................................................................................................................................264 Map 24: Strabo's Euphrates Route....................................................................................................................286 Map 25: Borderland Hubs.................................................................................................................................300 Map 26: Cities of the Expositio Totius Mundi et Gentium.............................................................................306 Map 27: The Cities of the Geographical Area of Mesopotamia......................................................................359 Map 28: Strabo's Assyria....................................................................................................................................362 Map 29: Adiabene and Environs.......................................................................................................................368 Map 30: Pliny's Syrian Lists..............................................................................................................................372 Map 31: Octavius' March..................................................................................................................................398 Map 32: Octavius' March, Satellite Image.......................................................................................................401 Map 33: Suggested Sites of Thapsacus.............................................................................................................404 Map 34: Proposed locations of Tigranocerta....................................................................................................414 Map 35: Kızıltepe Terrain.................................................................................................................................415 Map 36: Silvan Terrain......................................................................................................................................416 Map 37: Arzan Terrain......................................................................................................................................418 5/448 Acknowledgements First and foremost I would like to thank my dissertation committee: Claudia Moatti worked with me on every stage of the project; her valuable criticism shaped every part for the better. Kevin Van Bladel’s expertise on the languages, cultures and history of my study area was invaluable. Christelle Fischer-Bovet gave me meticulous and valuable comments on chapter drafts. Greg Thalmann and Ramzi Rouighi asked probing questions, threatened more, and ensured that I kept my readers in mind throughout. I could not have chosen a more supportive committee for my dissertation process, both academically and personally , and I cannot thank them enough. Thanks also to the faculty of the USC Spatial Sciences Institute, especially Karen Kemp, Darren Ruddell, Jennifer Swift and John Wilson. Without the initial training and encouragement of the faculty in the Department of Classics at the University of Canterbury , I would never have set foot on this arduous but rewarding journey . Thanks to Alison Griffith, Victor Parker, Enrica Sciarrino, Robin Bond, Graham Zanker and Tim Parkin: the love of Classical history and culture which they inspired still lives within these pages. Thanks also to the faculty of the Classics department at the Victoria University of W ellington: substantial portions of at least two chapters were written in their library and presented in their departmental seminar series. There is insufficient space to name all the other people who kept me sane throughout my graduate studies at USC, but special thanks go to Becky , Nick, and the rest of the Adams family; to my USC friends Christine, Sarah, John and Julie, Beau, Hannah, Kayla, Nicole, Tom, Mr and Mrs British Steel; to my Kiwi friends by birth, marriage or naturalisation: Rob and Becki, Jenn, Kirk, Olivia, Paula, Mash; and to my Los Angeles friends of a non-academic persuasion Morgan, James, Colin, Andrew , Gina and Jason, and Lindsey . Moreover, if I have shared pizza, beer, dice or archaeology with you over the past eight years, then I thank you too. Finally , thanks go to my parents, my sister and my nephew who keep me supplied with Skype calls at late and unexpected hours and ensure that Canterbury feels like a welcoming home whenever I return. To Glenda, my mother and biggest fan, my thanks and love for everything. This work is dedicated to my grandparents: Margaret, Jack, Max, Horace and Jean. 6/448 Introduction Geographic writing is often treated as an objective description of space, but the processes of selection, generalisation and aggregation that underlie geographic projects are subjective and ideological. The Mesopotamian frontier zone between the Roman and Iranian empires, that is, upper Mesopotamia, was the region between the Euphrates and Tigris rivers, south of the Taurus mountains and north of Parthian and Sasanian Babylonia. As the only frontier over which Rome faced an empire of similar size and power, this was arguably Rome’s most important frontier, yet its subjective and ideological representation in Roman literature is seldom studied. 1 My dissertation analyses geographic descriptions of that region written in the first four centuries CE, especially Strabo, Pliny the Elder, Claudius Ptolemy , and Ammianus Marcellinus as well as the anonymous Expositio Totius Mundi et Gentium. I examine the sources of their geographic knowledge to show the influence of Aramaic, Iranian, Hellenistic, and Roman cultural perspectives and scientific, administrative, and historical genres on their constructions and representations and how those representations fit within their conception of the Roman world as a whole. In other words, how the writers imagined the frontier and how they conveyed that imaginary to their audiences. I also address how these representations changed over the first four centuries of our era. The historical situation in the region is relevant to my work, especially as it is available for comparison to the descriptions given by the geographic writers, but my project does not aim to present a picture of historical “reality”, but rather of an ideological space of representation. 1 The Parthian empire (ca.247 BCE – 224 CE) was ruled by the Arsacid dynasty. The Sasanian Persian empire (224 – 651 CE) was ruled by the Sasanid dynasty. When writing of the dynasties themselves, I will use the dynastic adjectives (Arsacid and Sasanid); when writing of spaces, power structures, or the like, I will treat Parthian and Arsacid as synonymous and Sasanian, Persian and Sasanid and synonymous. I will occasionally have cause to refer to the Achaemenid Persian empire (ca.559 – 330 BCE), in which case I will use the dynastic and ethnic adjectives. If there is doubt, the meaning should be clear from the dates under discussion. I will describe these empires collectively as the Iranian empires. 7/448 This dissertation has three parts. In the first, I will show how the Roman imperial geographic writers constructed Mesopotamia as a space, what limits and names they assigned to it and how they represented local political changes in their geographical narratives. In the second part, I will discuss the past sources of knowledge used by the Roman imperial geographical writers and show how they used those sources to reflect contemporary political and ideological concerns. In the third part, I will consider the representation of mobility in the area. This is particularly visible in the economic sphere and in the representation of nomadic pastoralists. Throughout the work, I will use the idea of the “borderland” as a lens through which to examine the representations of Mesopotamia provided by the Roman imperial geographical writers. As I will explain below , this theoretical approach offers a productive way of thinking about the processes that occur in the spaces between large complex groups, such as the Iranian and Roman empires of my study period. Before I begin my analysis proper, there are several issues of background, methodology and terminology to consider. First, I will discuss the borderland as the theoretical framework which provides the overall structure to my project. Second, I will describe my study area geographically and chronologically . Third, I will provide a brief background on the various prior treatments that the region has received. The fourth section will concern geography: I will introduce several important terms and concepts in Roman geography and introduce my main sources. Other geographical sources will be introduced as they appear in the various chapters. Finally , I will outline those chapters and the overall structure of my analysis and argument in this work. 1. Borderlands In the early twenty-first century , borders are often conceived of as physically and legally fixed lines of demarcation. The ubiquity of geographic representations and information available in the modern world, particularly through the internet, has reinforced the idea that space is rigidly delimited by separating and 8/448 containing boundaries. Even in areas where borders are disputed, the existence of a border is seldom in dispute, rather where the border is to be drawn, and by whom. This mental framework is often problematic, even in modern times, and its application to historical situations can be misleading. 2 Social anthropologist, Fredrik Barth is concerned with the many ways in which boundaries can manifest. Barth defines a boundary as a conceptual construct impressed on the world to divide physical, social or cognitive spaces; 3 however, despite the cognitive imagery of boundaries as barriers, dividers, and stopping points, boundaries do not merely separate and contain, “impressing boundaries on the world creates affordances as well as limitations”. 4 In any sphere, some human activities will attempt to circumvent imposed boundaries. The great natural rivers to the north and east of the Roman Empire make convenient conceptual borders, but their cartographic convenience obscures a porous reality . Lee has persuasively argued that the idea that an armed limes formed an impenetrable wall surrounding the Empire is anachronistic. 5 Indeed, Lee marshals a significant body of evidence showing the movement of people and ideas back and forth across the Roman Empire’s northern and eastern frontiers. Yet, while boundaries are porous, they are still boundaries. Some actors may be limited by a boundary , while others will be afforded an opportunity . 6 As we shall see, both kinds of actor are evident in the Roman geographical sources. Another important assertion of Barth’s work is that boundaries are often analytical concepts imposed by outsiders. Maps of the Roman Empire often show a discretely bordered empire comprising a ring of polygonal spaces surrounding the Mediterranean. This gives the artificial impression of neatness and order, but the precision of that order is imposed by scholars for analytical purposes. The Roman imperial geographic writers also imposed their own boundaries for their own particular purposes. Defining those boundaries is a major goal of the first section of this 2 Elton (1996) 3. 3 Barth (2000) 17–20. 4 Barth (2000) 27. 5 Lee (1993). 6 Barth (2000) 27–30. 9/448 project and defining the purposes for which they were assigned is a major goal of the second section. In their discussion of colonial expansion in North America, Adelman and Aron attempt to disentangle the two notions of “frontier” and “borderland”. 7 They define a frontier as “a meeting place of peoples in which geographic and cultural borders were not clearly defined” and where “intercultural relations produced mixing and accommodation as opposed to unambiguous triumph”, while borderlands are “contested boundaries between colonial domains”. 8 Describing those North American borderlands, Adelman and Aron write: “In a pairing of the intercolonial and intercultural dimensions, differences of European rationales and styles come to the fore, as do shifts in those rationales and styles. Equally important to the history of borderlands and frontiers were the ways in which Indians exploited these differences and compelled these shifts, partly to resist submission but mainly to negotiate intercultural relations on terms more to their liking.” 9 In his response to Adelman and Aron, Evan Haefeli questions the distinction they draw between frontiers and borderlands. 10 Indeed, except for the presence of multiple colonial powers in a borderland and only one on a frontier, the experience of the local groups in the borderlands and frontiers described by Adelman and Aron seem to overlap considerably . Both cases are characterised by intercultural negotiation and accommodation and the exploitation of difference. More important is the relative capability of the colonial power to impose its will on local groups. A strong frontier presence grants the colonial power more capacity to impose by fiat, while a relatively weak frontier or a contested borderland allows local groups more room for autonomy , agency and opportunity . 11 These definitions of and distinctions between “frontiers” and “borderlands” are conditioned by the field within which they originated, North American history , and the 7 Adelman and Aron (1999). 8 Adelman (1999) 815–6. 9 Adelman (1999) 816. 10 Haefeli (1999). 11 For an example of accommodation and negotiation on a relatively weak frontier, see Shaw (1986), a study of the political relationships between Roman governors of the small plain area and heads of various peoples of the Mauretanian highlands commemorated in stone in the ancient Maghrib. 10/448 historical processes which operated in that period. Studies of Roman frontiers usually proceed from a military perspective, examining the region as part of a militarised frontier defence system. The huge field of “frontier studies” has produced a great many monographs, collected volumes and journals and is mostly concerned with the location of legionary bases, troop movements and questions of strategy . Freeman gives a brief summary of the origins of the field of Roman Frontier Studies in the nineteenth century and its voluminous scholarship fuelled by the large number of archaeological sites from the edges of the Roman empire (especially in Britain and Germany) as well as the considerable non-academic interest in such matters. 12 In the last thirty years, the field has expanded beyond the purely military or strategic elements to consider the effect of the army on the surrounding society . Isaac argues that Roman frontier policy was concerned with the use of the army for conquest, occupation and internal security , rather than for establishing a defensive perimeter. 13 Furthermore, he highlights the role of the provincial administration in fostering urbanisation, including, but not limited to, the foundation of veteran colonies. The role of the army in road building is of particular importance to the present project. 14 In a particularly important work, Whittaker examines the Roman ideology of the frontier and compares it to Roman frontier practice across the physical breadth and chronological depth of the Empire. 15 He argues that at no point is there any sign that Roman ideology saw a fortified defensive system as marking the limit of their operational space on the Eastern frontier. Almost contemporaneous with Whittaker, Elton considers the frontier as a series of overlapping zones representing different groups performing different activities. 16 To construct a simplistic example, the frontier in a particular area might 12 Freeman (1996). For example, there are multiple JRA volumes, journals and conference publications on the subject of Roman frontier studies. 13 Isaac (1990). 14 Although mostly concerned with southern Syria, the collected papers in Isaac (1998) further develop these themes. 15 Whittaker (1994); Whittaker (2004). 16 Elton (1996). 11/448 comprise the military boundaries of a Roman legionary camp, the linguistic boundaries of a local people and the economic boundaries of a particular town. On this basis, Elton examines the establishment and operation of frontier zones and the activity (particularly commercial and military) of their inhabitants across the Empire. While these studies have expanded beyond the military focus, that sphere is still particularly evident in the goals and methods of much of the field’s scholarship. As part of his study of the northern frontier regions of the Assyrian empire, Archaeologist Bradley Parker translates the historical definitions constructed in the context of North American history into a general model for considering the types of borders present in an ancient borderland. In his 2006 article, “Toward an Understanding of Borderland Processes”, Parker defines the terms boundary , border, frontier and borderland. 17 A boundary is “that which serves to indicate the bounds or limits of anything”. 18 Borders and frontiers are types of boundary , the former being a spatially fixed dividing line, and the latter a zone of overlapping boundaries. In this formulation, a borderland is a geographic space around or between political or cultural entities where circumstances or processes arising from the interaction of boundaries create borders or frontiers. Parker’s work on borderlands builds upon Hugh Elton’s study of the Roman frontiers. 19 Elton stresses the necessity of considering the Roman frontiers as overlapping series of zones. It is important to stress the two factors at play here: 1) that the frontier be considered as a zone, not a line, 2) that it be considered as a number of overlapping zones, rather than a single zone. Nevertheless Elton’s work emerged from the field of Roman frontier studies and was concerned primarily with military spaces and the interaction of soldiers and locals within those spaces. Parker’s development of Elton’s concept of overlapping zones explicitly reduces the military aspect of the frontier to one of several aspects. 17 First described in Parker (2002) and fully elaborated in Parker (2006). Political boundaries will be discussed further in Chapter 2. 18 Parker, (2006) 79, accepts the Oxford English Dictionary’s definition. 19 Elton (1996). 12/448 Each of the boundaries that interact to create a given borderland can be described on what Parker calls the Continuum of Boundary Dynamics. At one end of the continuum are borders, static and restrictive, and at the other are frontiers, porous and fluid. There are many types of boundary that could exist in a borderland situation and be represented on this continuum. Parker proposes that they can be divided into five categories, which he calls boundary sets: geographic, political, demographic, cultural, and economic. 20 The geographic boundary set includes topographic features, the physical character of the landscape, climate, ecological boundaries, and natural resources. The political boundary set includes political, administrative and military boundaries. The demographic boundary set includes ethnic, population density , health and gender boundaries. The cultural set includes linguistic and religious boundaries, as well as those between different material culture assemblages. Finally , the economic set includes the extraction, transportation and production of materials, commodities and goods. All of the boundaries that will be considered in this project fall into one or more of these categories. Chapter one is largely concerned with the representation of geographical boundaries (such as topography), chapter two with political and administrative boundaries (kingdoms and provinces), chapters three and four with cultural boundaries (the ideological distinctions between Roman and Iranian space), and chapter five with economic boundaries (productivity and transportation). 21 Examining the circumstances and processes that contribute to these categories and the interactions between these boundary sets and their elements through time is an important tool for thinking about and characterising the processes which take place in a given borderland. 22 A handful of studies have examined these processes in the context of the Mesopotamian borderland. Lee examines the role of information flow 20 The lists that follow are from B.J. Parker (2006) 82. 21 The Roman authors give very little information about the Mesopotamian borderland which would fall within the demographic boundary set. 22 “Borderland processes can be defined as the dynamic interaction within and between boundary sets as their characteristics (i.e. static, restrictive, porous, fluid) and the nature of their interconnections vary through time.” B.J. Parker (2006) 94. 13/448 across frontiers; both general background knowledge about distant people and places, and current information of immediate relevance to decision-making. 23 Most importantly for my project, Lee shows that, contrary to some previous scholarship, frontiers did not pose a barrier to movement. While Lee discusses the flow of information across boundaries, Graham analyses the flow of information from the frontiers to explore how that information altered Roman frontier ideology . 24 In a recent article, Moatti describes the increasing desire in the later empire to regulate the movement and activities of foreign merchants and certain other classes of travellers. 25 In particular, she looks at the role of the treaties between the Sasanians and Rome in limiting trade to particular Mesopotamian cities, most famously Nisibis. 26 These empire-wide developments were reflective of a growing Roman awareness of their ability to legally define and regulate the limits of their empire. 27 The late imperial focus of these studies raises an important point of chronological change. There is no doubt that the nature and rigidity of the frontier changed from a broad, largely symbolic zone between the Parthians and the Republican dynasts to a more clearly demarcated and fortified border after Diocletian in particular. For this reason, many studies on the frontiers or the region generally mark the end or beginning of their study period around the late third and early fourth centuries CE. 28 At a closer scale, well-preserved urban sites invite examination of cultural boundaries. Cities like Dura Europus, Edessa, Palmyra, and Zeugma, all show evidence of the continuation of local cultures to varying degrees. 29 Such cities often provide a fertile space for consideration of boundaries. Some of these are evident in the geographical sources, but usually a lack of literary evidence means that a close examination of 23 Lee (1993). 24 Graham (2006). 25 Moatti (2011). 26 Moatti (2011) 13–17. 27 Moatti (2011) 27–8. 28 For example: Millar (1993); Lee (1993); Sartre (2005); Graham (2006); Edwell (2008). 29 Dura Europus: Wharton (1995) 15–63; Dirven (1999); Pollard (2004); Sommer (2006); Pollard (2007). Edessa: Segal (1970); Drijvers (1980); Drijvers and Healey (1999); Ross (2001). Palmyra: Drijvers (1976); Teixidor (1984); Will (1992); Gawlikowski (1994); Sartre (1996); Gawlikowski (1996); Gawlikowski (2003); Yon (2004); Smith (2013). Zeugma: Kennedy (1994); Abadie-Reynal et al. (1998); Kennedy (1998); Desreumaux et al. (1999); Gaborit and Poccardi (2000). 14/448 material remains is required instead. Such examinations are beyond the scope of this project. All of these cities will appear in my discussions, but the Roman imperial geographic writers seldom give sufficient information on any one city to analyse them in this way . Finally , important recent work on the role of geography in the formation of historical, cultural and economic regions will inform my project. While Rome was a Mediterranean empire, Mesopotamia was one of a few areas where imperial control extended beyond that geographic, economic and cultural zone. Many scholars have demonstrated the coherence of the Mediterranean as a region in various periods, especially Horden and Purcell's ecological and geographical approach. 30 The degree to which Roman control somehow overcame geography to extend that region to include inland areas or whether those areas are better seen as coherent regions themselves is particularly relevant to the question of the relationship between the frontier and the whole empire, in reality or in representation. 31 The Mesopotamian and Armenian frontiers were arguably Rome’s most important border. Certainly the only border across which she faced a political entity of similar power. Many scholars have written about this region as a frontier but none have approached that study from a geographical perspective. Moreover, while recent scholarship has nuanced the rather simplistic view of Roman frontiers as static defensive lines that persisted as recently as the 1990s, much further work remains to be done to understand the role that mobility played in shaping the region as a borderland space and as a space within the Roman geographical imagination. This dissertation is informed by all these perspectives and the regional and urban studies in particular provide valuable comparanda for the representations of those spaces in the works of the geographical writers. Research on the frontiers in general continues to expand our picture of individual, collective and administrative practice on the frontiers. Yet while Whittaker in particular has opened the 30 Horden and Purcell (2000). 31 See for example Batty's examination of Rome's contacts with and intrusions upon the “Pontic-Danubian” region spanning the plains west and north of the Black Sea: Batty (2007). 15/448 relationship between ideology and the frontier, the role of geographical representations in creating and reinforcing that ideology is underexplored. 2. Study Area My study area comprises the area of northern Mesopotamia contested by the Roman and Iranian empires over the first four centuries CE. The core of the area is bounded by the Taurus mountains to the north, the Tigris to the east and the Euphrates to the south and west. However, I also include, to the west, Commagene in the Taurus foothills and the pastoral tribes between the urban centres of the Roman province of Syria and the Euphrates, and, to the south, the northern limits of the Arabian steppe area dominated by Palmyra and its agents who controlled and in fact constituted a network of desert trade routes between Syria and Babylon. Thus, the study area includes the territories that would be encompassed by the Roman provinces of Osrhoena and Mesopotamia, as well as the adjacent regions of Commagene and Adiabene whose histories were intertwined with the frontier in a fundamental way . These areas exhibit a considerable degree of geographical and cultural continuity . 32 This area is usually examined as part of studies on the Roman Near East generally . Inasmuch as these regions shared a history and culture, this approach is justified. However, the distribution of the evidence naturally favours the well-preserved and researched cities of the Syrian coast. Millar and Sartre exemplify this style with the wide scope and focus on cultural change of their handbooks. Both utilise an exhaustive array of traditional historical sources to address the profound changes that Roman rule had on the communities of the Near East between Anatolia and Egypt. 33 Sartre organises his material thematically , examining how the 32 This cultural continuity was an ancient perception, see, for example, Strabo 16.1.1 where he describes Assyria as a unitary region. Philostratus (V A 1.16) refers to a local myth about Antioch on the Orontes as “Assyrian”; see also V A 1.19. For examples of artistic similarities between Hierapolis and Hatra, see Kropp (2013), and cultural similarities between Hatra and Palmyra, see Yon (2013). 33 Millar (1993); Sartre (2001); I have consulted primarily the updated English translation which treats only the Roman material (c. 69 BCE to 272 CE): Sartre (2005). 16/448 Syrian cultural landscape was changed by the arrival of Hellenism, then how Rome appropriated, organised, and was in turn influenced by that new landscape. Millar takes a regional approach, within which he addresses similar questions of interaction between Roman, Greek and local Aramaic cultures. One of his concerns is the interaction between the sedentary space of Roman-controlled Syria and the non-sedentary space in the steppe beyond. As Near Eastern archaeologist, Ball approaches the subject from two directions quite distinct from the traditional classical perspectives of Millar and Sartre. 34 The first is an Iranian perspective, explicitly challenging the Eurocentrism of traditional classical approaches to the study of the Roman Near East. The second is an art-historical perspective, including extensive comparison and analysis of artistic and architectural evidence. Moreover, Ball explicitly addresses the influence of the Near East on Rome. Another recent work with a partially art-historical approach is Butcher. 35 Like Millar and Sartre, Butcher particularly examines cultural changes and the Hellenisation of the Near East. One of the few regional monographs to explicitly study Mesopotamia is that of Edwell. 36 His study area is focused on the Euphrates itself, roughly between Samosata and Dura-Europus, but including other areas pertinent to the frontier such as Palmyra and Hatra. Edwell critically evaluates many Mesopotamian archaeological sites, especially Dura-Europus, and the Sasanian royal inscriptions as they pertain to Shapur’s campaigns of the mid-third century . While the spatial and chronological dimensions of Edwell’s study are very similar to those of my own project, his focus is firmly on the political and military aspects of the region. Mesopotamia and the middle-Euphrates have also been the subject of a number of broader archaeological studies which shed light on the role of mobility the region. Several recent articles by Wilkinson have synthesised the dozens of archaeological survey projects which have examined the landscape of northern Mesopotamia. 37 The interrelated advancements in the technological power, popular awareness 34 Ball (2000). 35 Butcher (2003). 36 Edwell (2008). 37 Wilkinson et al. (1994); Wilkinson (1995); Wilkinson (2000); Wilkinson et al. (2005). Wilkinson’s primary 17/448 and affordability of remote sensing technology in particular have proved invaluable in this area. Kennedy and Riley's assembly of aerial photography , including several pioneering studies in the early twentieth century provided new literal and figurative viewpoints on regional sites. 38 More recently , Comfort, Abadie-Reynal and Ergeç have used satellite imagery to study crossing points on the Euphrates and show the multiplicity of available routes connecting Syria with Mesopotamia. 39 In more traditional research, Gaborit and Leriche provide a thorough compilation of sites on the middle-Euphrates using evidence from geographical writers from the Hellenistic to Byzantine periods, as well as travellers' accounts from the seventeenth to early- twentieth centuries, and recent archaeological work. 40 “Mesopotamia” and the Mesopotamian Borderland The term “Mesopotamia” has a varied and complicated history . In modern scholarship, particularly in the field of archaeology , the entire area between and around the Euphrates and Tigris rivers, from the Taurus to the Persian gulf, receives the name “Mesopotamia.” The northern half, where settled cultivation did not spread far beyond the Taurus or the rivers and where much of the land between those cultivated zones was desert, is known as Northern or Upper Mesopotamia. The southern half, where a much greater proportion of the area between the rivers was cultivatable, or was made so by canals, is called Southern or Lower Mesopotamia. The spectacular sites and artefacts of Babylon, Sumer and Ur have firmly attached to the term “Mesopotamia” to this southern region in the popular imagination. Because of this modern association, archaeologists of northern Mesopotamia often use the term the Jazira (from the Arabic al-Jazira, the Island) for the dry , northern region. Like “Mesopotamia”, Jazira refers to the physical characteristics of the space as surrounded by the waters of the Euphrates and Tigris. Also like Mesopotamia, it is a term with several interest lies in the Assyrian empire, but his discussions and the survey evidence he presents have a much broader chronological scope which includes the Roman period. 38 Kennedy and Riley (1990). 39 Comfort et al. (2000); Comfort and Ergeç (2001). 40 Gaborit and Leriche (1998). 18/448 different spatial definitions in different periods of Islamic rule. As we shall see in Chapter 1, Greek, Hellenistic and Roman geographic writers used “Mesopotamia” as a geographical term to refer to the land between the Tigris and Euphrates rivers which lay south of the Taurus and north of Babylonia (the area between the rivers, labelled “Mesopotamia” in Map 1: The Mesopotamian Borderland). Their conception of the northern and southern boundaries of the space varied. The term never extended into the Anatolian plateau, but sometimes included the southern foothills of the Taurus Mountains. While it was occasionally used to refer to the entire zone all the way south to the Persian gulf, that southern part of the region was more commonly referred to as Babylonia. Moreover, “Mesopotamia” was also at times an administrative term which referred to a smaller area within that geographical space. The first part of this dissertation will examine these delimitations and denominations in detail. For this reason, when I have cause to refer to my study area as a whole, I will refer to it as “the Mesopotamian borderland” (roughly the highlighted area of Map 1: The Mesopotamian Borderland). By this I mean the borderland between the Roman and Iranian empires which lay in Mesopotamia. 41 Referring to this space as the Mesopotamian borderland opens it to more than the physical area between the rivers, and alleviates the ambiguity which surrounds the term “Mesopotamia” and its various uses throughout time, space and genre. In using the term, I aim to consider an area broader than “Mesopotamia” and narrower than the Romano-Iranian borderland. As Map 1 shows, I explicitly include areas to the east of the Tigris and west of the Euphrates. Rivers act more as conduits than barriers and often inextricably link conditions on either side of them. For example, as I will show in Chapter 2, the history of Roman expansion into and representation of Osrhoene is enriched by considering it in parallel with Roman expansion into and representation of Commagene. The Mesopotamian borderland is distinct from other borderlands between 41 Not a borderland or edge internal or adjacent to Mesopotamia as a geographical space. 19/448 those empires, in particular that in Armenia, which is explicitly not part of my study area. While Armenia was a major theatre of conflict for almost as long as the Roman empire had territorial interests in Asia and disputes in that region often heralded military action south of the Taurus, quite different geographical and cultural conditions existed there. 42 The First Four Centuries Chronologically , my study area is defined by Roman victories in the first century BCE and Sasanian victories in the fourth century CE. It begins with the arrival of Roman power in the Near East, especially the 42 For a persuasive argument that Armenia and Mesopotamia should be considered together, see De Jong (2013). However, his argument applies mainly when the object of analysis is Parthian culture. 20/448 Map 1: The Mesopotamian Borderland direct control of kingdoms and cities brought about by the campaigns of Lucullus and Pompey against Mithridates of Pontus and Tigranes of Armenia. This highlights the Roman and geographical perspectives of my project. As I will argue in Chapters 2 and 3, the geographical sources which survive from the Roman imperial period must be treated as products of their time. While they are full of Hellenistic information, they were constructed within the social and political conditions of the Roman empire and cannot be entirely divorced from that context. My period ends with Jovian’s defeat by Shapur II in 363 CE and the subsequent treaty which ceded Nisibis to the Sasanian Persian empire. Between the Antonine Itinerary , the Peutinger Table, Ammianus Marcellinus, and the Expositio Totius Mundi et Gentium, the fourth century was comparatively rich in geographical texts. Although Strabo and Pliny provide our two most detailed narrative descriptions of the area and thus are frequently the focus and starting point of the various subjects I examine within this investigation, extending the chronological scope to include these fourth-century texts offers an fascinating contrast between the earlier and later texts and allows a more varied selection of perspectives on the Mesopotamian borderland. In particular, it allows us some perspective into the changes that took place in the third century , for which we mostly lack good evidence, in geography as in history . Moreover, the treaty of 363 can be seen as a conceptual end to Roman expansion in the east. Thereafter, Roman and Persian power were largely in balance and the boundaries between them became more rigid, particularly in the military and ideological spheres. 43 At the scale presented by the Roman imperial geographical writers, borderland interactions are most visible when the borderland is more fluid and mobile. In this state, the effect of imperial policies and geographical changes on smaller polities and the description of those polities can be seen. 43 Dignas and Winter (2007) provides a relatively brief but thorough examination of relations between the two empires between the third and seventh century. 21/448 Physical Geography Northern Mesopotamia can be divided into three broad zones: a fertile shelf near the Taurus mountains, a barren rocky desert plateau, and the river valleys of the Euphrates, Tigris, and their tributaries. Dryland farming, dependent on natural rainfall, is possible on the fertile shelf. 44 The large areas of space in northern Mesopotamia compensated for the low crop yields relative to southern Mesopotamia. 45 As one proceeds south, farming becomes increasingly marginal until the land gives way to desert where sparse grazing, wells and oases provide sustenance for goat and camel pastoralists. In the third millennium BCE, settlement in this area was concentrated in large circular walled settlements dot this area, but these were replaced by a dispersed settlement pattern by the first millennium. 46 In the Hellenistic and Roman periods, most of the major cities of the region were in the northern zone, among the headwaters of the Balikh and Khabur rivers. These rivers flow south into the Euphrates offering easily irrigated strips of land for cultivation, and convenient routes through the arid steppe around them. As will be seen, this environment played an important role in the dynamics of the borderland. A Note on the Maps This dissertation contains a number of maps which I constructed to aid the reader in their comprehension of my arguments. While visualisations of the spatial relationships help us think about the spaces we examine, it is important to remember that just like geographical writing, cartography presents an image which the creator designs. 47 I constructed the maps using the ArcGIS 10 package. The basemaps were provided by Bing Maps through ArcGIS Online, accessed through ArcMAP 10. 48 The maps vary in scale, 44 Wilkinson (1994) 484–85; Wilkinson (1995) 144–45; Wilkinson (2000) 222. 45 W eiss (1986). “The archaeological reconstruction of land-use patterns within several ecological zones if the Jazira provide support for this model of large-scale prairie-type agriculture, but with an added emphasis on differential levels of land-use intensity dependent upon location.” Wilkinson (2000) 222. 46 Wilkinson (1995) 139–59; Wilkinson (2000) 236, 239. It is unclear why , but the leading theories are Assyrian resettlement policies or the spontaneous movement of Aramaeans. 47 Smith (2005) 832. 48 Bing Maps: http://www .bing.com/maps. 22/448 but are consistent in map projection (GCS WGS 1984) throughout for ease of comparison between maps in the document. 49 The point data for the ancient sites was gathered from a variety of online sources including the Pleiades Project and Wikipedia. 50 The Pleiades Project is particularly cautious, and often gives estimated areas for sites where any doubt exists, or sometimes even when a site if well known. Wikipedia coordinates were often more useful. All coordinates were checked against published maps, usually those of Sartre, Millar, Edwell or the Barrington Atlas as well as close range examination using Google Maps or Acme Mapper. 51 The resolution of publicly accessible satellite imagery is usually sufficiently detailed to make out ancient ruins and compare them to published site plans. In general, I am confident in the placement of the sites I locate; where doubt remains, I indicate so in the text. By and large, the geophysical terrain features of the ancient landscape remain today . However, the main problem with constructing ancient maps using modern terrain data is the issue of water levels. For inland parts of modern Turkey , Syria and Iraq, this problem arises from the numerous dams which have been built on the Tigris and Euphrates rivers. Several modern hydro-electric and irrigation schemes have produced large lakes on the Euphrates. Accordingly , the modern boundaries and courses of the water bodies shown on these maps may not correspond to the ancient boundaries and courses of those bodies. 3. Geography An important goal of this dissertation is an analysis of the way in which the description of a region can reveal ideological structures. As Claude Nicolet notes, “geography should not be understood as a reality 49 GCS WGS 1984 is a convenient standard, which is used by the GPS system: http://blogs.esri.com/esri/arcgis/2008/07/18/gcs-wgs84-why-should-you-care-about-it/ 50 http://pleiades.stoa.org/home ; https://www .wikipedia.org/. 51 Millar (1993); Talbert and Bagnall (2000); Sartre (2005); Edwell (2008); http://mapper.acme.com/; http://www .google.com/maps. 23/448 but as a representation of a reality .” 52 The idea that space is constructed and ideological is fundamental to my analysis. Harvey describes the development of these ideas in philosophical thought, culminating in Lefebvre's La Production de l’espace (1974) in which Lefebvre develops earlier tripartite theories of space (as absolute, relative and relational, or organic, perceptual and symbolic) as his own division of spaces into “material space (the space of experience and of perception open to physical touch and sensation); the representation of space (space as conceived and represented); and spaces of representation (the lived space of sensations, the imagination, emotions and meanings incorporated into how we live day-by-day)”. 53 Said called geography “a socially constructed and maintained sense of place” to emphasise its proscriptive power. 54 Several studies have examined the construction and manipulation of spatial representation in an ancient context. 55 Nicolet has examined the ideological construction of space in a Roman context, specifically the relationship between power and geography in the early Roman Empire. 56 Geographic thought flourishes when times of conflict and conquest cause re-evaluations of the world. 57 Nicolet shows how Roman imperial ideologies at the time of Augustus helped to construct and were in turn constructed by Roman conceptions of space. This relationship between imperialism and geography has been extensively explored by modern geographers, as Clayton's examination of recent research on imperial and colonial geographies in cultural geography shows. 58 As in other areas of classical research, postcolonialist ideas about the ideological character of knowledge and the organisation of knowledge can be fruitfully applied to geography . 59 The conceptional organisation and representation of the empire in the geographical texts play into Roman conceptions of their 52 Nicolet (1991) 3. 53 Harvey (2006) 271–80, the quote is from p.279. For an expanded description of Lefebvre's conception, see Thalmann (2011) 22–3. I am consulting the English translation, Lefebvre (1991). The work of Soja is also relevant Soja (1996). 54 Said (2000) 180. On the culturally and politically contingent nature of maps, see Rihll (1999) 101. 55 Romm (1992) on wonder literature; Parker (2008) on Roman India; Thalmann (2011) on Apollonius of Rhodes. 56 Nicolet (1991). See also Ando (2000) 320–29. 57 Clarke (1999) 69. 58 Clayton (2004). 59 For the application of such ideas to a variety of imperial genres, see the papers collected in Konig and Whitmarsh and especially their introduction: Konig and Whitmarsh (2007). 24/448 empire and its place in the world. Lewis and Wigen call this metageography: “Every global consideration of human affairs deploys a metageography , whether acknowledged or not. By metageography we mean the set of spatial structures through which people order their knowledge of the world: the often unconscious frameworks that organize studies of history , sociology , anthropology , economics, political science, or even natural history .” 60 The representation of the border regions of the Roman empire plays into this broader imperial metageography . Geography in the Roman Empire One of the goals of this project is to illuminate the practice of geographical writing in the Roman Empire. Before doing so, it is necessary to make a few general comments on geographical terms and concepts that I will employ throughout this work. These are my use of the term “Roman imperial geographic writers”; the distinction between scientific and descriptive geography; the scope of the various categories of what we call geography today , including topography , chorography and geography; and the important role of linear conceptions and description in ancient geographical thought. 61 Roman Imperial Geographic Writers Not all of the authors examined herein would have considered themselves “geographers” and few of the works discussed were called “geographies”. Accordingly , I have avoided using the term “geographer” as my project addresses geographical thought beyond the usual classification of that genre. In particular, Clarke has problematised the boundaries between the genres of geography and history . 62 While the relationship between history and geography is usually seen as analogous to those between time and space, or past and 60 Lewis and Wigen (1997) xi. For Roman ideas about the continents as they relate to the empire, see Clarke (1999) 210–228. 61 Nicolet collapses the distinctions between scientific and descriptive and linear and planar modes, and would have the two types of geography be scientific and linear (periplus), Nicolet (1991) 58–9. However, as will be seen, non-scientific geographers like Strabo and Pliny were able to conceive of and describe the world in a planar sense. 62 Clarke (1999). 25/448 present, Clarke argues that ancient writers did not draw such a hard line between the two genres. 63 Strabo certainly wrote works in both modes, and even within the text of his Geography , there are frequent “historical” moments. The same can be said mutatis mutandi for Herodotus, Xenophon and Ammianus Marcellinus. 64 Similar problems emerge in classifying the work of Pliny the Elder. Thus, my body of evidence includes some authors traditionally considered geographers and some traditionally considered historians, as well as some texts that might not be considered within either genre. Regardless, they are all geographic writers of some sort and their geographical passages may generally be described using the same categories as geographers, narrowly defined. Scientific and Descriptive Geography An important internal distinction within ancient geography is that between the scientific and descriptive modes. 65 Scientific geography is concerned with precisely recording distances and locations for the features of the earth's surface. This was the style of geography practised by the great Hellenistic geographers Eratosthenes, Hipparchus and Posidonius. Ptolemy's comprehensive record of the coordinate locations of geographic features is the classic example for the imperial period. According to Ptolemy , geography requires measured and proportional positions rather than artistically described qualities. 66 Most of the extant geographic writers of the imperial period are less concerned with scientific precision than Ptolemy and his scientific forebears. Descriptive geography aimed to describe landscapes through narrative. It often included ethnographic information as a fundamental part of the descriptive process. 67 For example, the geographical writing of Strabo aimed at diverse and detailed narrative description which would be of 63 Clarke (1999) 4–22. 64 Sundwall (1996); Harrison (2007); Roy (2007). 65 Dueck (2010) 236–7; Dueck (2012) 20–98. 66 Ptol. Geog. 1.1. For a description of scientific methods of establishing coordinate locations using shadows, see Evans (1998) 59–63, 99–100, 102–3; Dueck (2012) 90–98. 67 I would include chorography within the rubric of descriptive geography , following Romer that the difference between chorography and geography was simply scope, Romer (1998) 4–5. 26/448 practical use. 68 Despite these differences in aims and methods, descriptive geography is often criticised using the criteria of scientific geography . 69 However, Romm argues that “for the ancient Greeks, and to a lesser degree for Romans as well, geographia represented a literary genre more than a branch of physical science” relying more on narrative rather than simple description. 70 Indeed, Strabo notes that measurements are a necessary but difficult (περισκελὲς) part of geography , which must be endured (ὑπομένειν). 71 The Scope of Geography The level of detail at which ancient geographical writers considered their works to operate has a significant bearing on the way they selected material for inclusion and aggregated areas which they considered homogeneous at the desired level of representation. Ancient authors conceived of geography and chorography as slightly different geographical projects. Both Ptolemy and Strabo understood that descriptive geography was based on a process of selection and qualitative judgement. The difference between geography and chorography that was essentially one of scale. Geography sought to describe the entire world, continents, coastlines, regions and provinces, while chorography focused on a region and provided far more details. 72 Ptolemy opens his Geography with a discussion of the difference 68 Strabo 1.1.16; 2.5.13; Dueck (2000) 53–62. In the first line of his work, Strabo (1.1.1) calls geography a πραγματεία: “τῆς τοῦ φιλοσόφου πραγματείας εἶναι νομίζομεν, εἴπερ ἄλλην τινά, καὶ τὴν γεωγραφικήν, ἣν νῦν προῃρήμεθα ἐπισκοπεῖν.” (I think that geography , which I propose to examine, is, as much as any science, the business of philosophy). All translations are my own unless otherwise indicated. 69 For example, Goodyear on Pomponius Mela’s De chorographia: “[it] has not won the approval of geographers, though Pliny the Elder, hardly a discriminating critic, seems to have taken in seriously. … [Mela’s] worst fault is that he supplies no measurements.” Goodyear (1982) 667. 70 Romm (1992) 3. See also Sundwall (1996) 622–23. 71 Strabo 14.1.9: “μικρὸν δὲ πλέον τὸ ἀπὸ Μιλήτου εἰς Ἡράκλειαν ἐγκολπίζοντι, εὐθυπλοίᾳ δ᾽ εἰς Πύρραν ἐκ Μιλήτου τριάκοντα: τοσαύτην ἔχει μακροπορίαν ὁ παρὰ γῆν πλοῦς. ἀνάγκη δ᾽ ἐπὶ τῶν ἐνδόξων τόπων ὑπομένειν τὸ περισκελὲς τῆς τοιαύτης γεωγραφίας.” (But the voyage from Miletus to Heracleia, including the sinuosities of the gulfs, is a little more than one hundred stadia, though that from Miletus to Pyrrha, in a straight course, is only thirty—so much longer is the journey along the coast. But in the case of famous places my reader must needs endure the dry part of such geography as this.) Loeb trans. 72 To this could be added topography , very closely focused examinations of particular towns or cadastral land divisions around them, but the Roman imperial geographical writers do not consider this topic: Nicolet (1991) 4; Dilke (1971). 27/448 between geography and chorography: Ἡ γεωγραφία μίμησίς ἐστι διὰ γραφῆς τοῦ κατειλημμένου τῆς γῆς μέρους ὅλου μετὰ τῶν ὡς ἐπίπαν αὐτῷ συνημμένων· καὶ διαφέρει τῆς χωρογραφίας, ἐπειδήπερ αὕτη μὲν ἀποτεμνομένη τοὺς κατὰ μέρος τόπους χωρὶς ἕκαστον καὶ καθ’ αὑτὸν ἐκτίθεται, συναπογραφομένη πάντα σχεδὸν καὶ τὰ σμικρότατα τῶν ἐμπεριλαμβανομένων, οἷον λιμένας καὶ κώμας καὶ δήμους καὶ τὰς ἀπὸ τῶν πρώτων ποταμῶν ἐκτροπὰς καὶ τὰ παραπλήσια· τῆς δὲ γεωγραφίας ἴδιόν ἐστι τὸ μίαν τε καὶ συνεχῆ δεικνύναι τὴν ἐγνωσμένην γῆν, ὡς ἔχει φύσεώς τε καὶ θέσεως [καὶ] μέχρι μόνων τῶν ἐν ὅλαις καὶ περιεκτικωτέραις περιγραφαῖς αὐτῇ συνημμένων, οἷον κόλπων καὶ πόλεων μεγάλων ἐθνῶν τε καὶ ποταμῶν τῶν ἀξιολογωτέρων καὶ τῶν καθ’ ἕκαστον εἶδος ἐπισημοτέρων. 73 For Ptolemy , the primary difference between geography and chorography is the level of selection required by the practitioner of each. Geography is more selective, chorography less so. This is because chorography aims to give an impression of a part, while geography aims to give an impression of the whole. 74 Ptolemy goes on to define the field within which he sees himself in terms of scientific geography; geography falls within the realm of mathematics, chorography within the realm of the artist. 75 Geography looks at position, chorography at quality . Not all geographic writers would have agreed with Ptolemy’s classification. While Strabo insisted that a good geographer master mathematics, physics and a host of other sciences, these skills provide a background to the practical purpose of his work. Strabo periodically gives mathematical data and spends lengthy passages criticising details of measurement in his predecessors’ work, but in the bulk of his narrative, he deals in description. Strabo recognised the importance of astronomy and mathematics but described the work of the geographer as a more holistic philosophical project. 76 Strabo 73 Ptolemy 1.1.1-3: “Geographia is an imitation through drawing of the entire known part of the world together with things that are, broadly speaking, connected with it. It differs from chorographia in that chorographia, as an independent discipline, sets out the individual localities, each one independently and by itself, registering practically everything down to the least thing therein (for example, harbors, towns, districts, branches of principle rivers, and so on), while the essence of geographia is to show the known world as a single and continuous entity , its nature and how it is situated, [taking account] only of the things that are associated with it in its broader, general outlines (such as gulfs, great cities, the more notable peoples and rivers, and the more noteworthy things of each kind).” Translation adapted from Berggren and Jones (2000). Berggren and Jones translate geographia as “world cartography” and chorographia as “regional cartography”, (2000) 57, n.1. 74 Ptol. Geog. 1.1.2-3. 75 Ptol. Geog. 1.1.4. 76 Strabo 1.1.13-20; 2.5.1-2. 28/448 focused on the political and the practical uses of geography , although he realised that there was a limit to the amount of detail one could include in such a work. καθάπερ τε καὶ ἐν τοῖς κολοσσικοῖς ἔργοις οὐ τὸ καθ᾽ ἕκαστον ἀκριβὲς ζητοῦμεν, ἀλλὰ τοῖς καθόλου προσέχομεν μᾶλλον εἰ καλῶς τὸ ὅλον, οὕτως κἀν τούτοις δεῖ ποιεῖσθαι τὴν κρίσιν. κολοσσουργία γάρ τις καὶ αὕτη, τὰ μεγάλα φράζουσα πῶς ἔχει καὶ τὰ ὅλα, πλὴν εἴ τι κινεῖν δύναται καὶ τῶν μικρῶν τὸν φιλειδήμονα καὶ τὸν πραγματικόν. ὅτι μὲν οὖν σπουδαῖον τὸ προκείμενον ἔργον καὶ φιλοσόφῳ πρέπον, ταῦτα εἰρήσθω. 77 Because of the size of his chosen topic, Strabo had to be selective and omit details; “sacrificing detail for overall impression”. 78 Throughout his work, Strabo uses the term geographia (τὰ γεωγραφικά) to describe his project. But Strabo's colossal work (κολοσσουργία) is significantly more detailed than those of his predecessors, verging on chorography . He uses the term chorographia (τὰ χωρογραφία) to describe at least a part of his work in the extremely detailed portion on Greece. 79 In fact, this expansion of the purview of “geography” to include a framework of scientific geography hung with chorographic detail is Strabo's major innovation. 80 The role of selection of material in the representation of space is a recurring theme in my project. Both Ptolemy and Strabo recognised the necessity , or desirability , of choosing their material to suit the image of the space which they wanted to convey . The metaphor of an artist creating a representation of reality is particularly apt. 77 Strabo 1.1.23: “Just as in with colossal statues we do not seek precision in each individual part, but rather consider the general effect and whether the statue as a whole is pleasing, so should this book of mine be judged. For it too is a colossal work, which shows only large and wholes things, except where even something small is able to stir the interest of the studious or the practical man. I have said these things to show that the present undertaking is worthy of serious attention and suitable for a philosopher.” 78 Pothecary (2005) 6–10, the quote is from 9. 79 In his discussion of Elis: Strabo 8.3.17: “παραπλησίως δ᾽ ἂν καὶ τὰ περὶ τοῦ Πύλου διαπορούμενα τύχοι τῆς προσηκούσης διαίτης ἐπελθοῦσι μικρὸν ἔτι τῆς χωρογραφίας μέχρι τοῦ Πύλου τοῦ Μεσσηνιακοῦ” (And similarly , too, the puzzling questions raised in regard to Pylus may find an appropriate solution when, a little further on in my chorography , I reach the Messenian Pylus.) Loeb trans. Dueck, (2000) 155, questions to what extent he intended to apply the term chorographia to his entire work, or a portion. 80 Dueck (2000) 156. 29/448 Linear Conceptions of Space Another important distinction is that between planar and linear geography . Roman imperial geographic writers conceived of and described space in both two-dimensional and one-dimensional ways. Descriptive geographers like Strabo provide ample evidence for the ability of Romans to conceive of space in two-dimensional terms, but the role of linear descriptive techniques in Classical thought has been recognised and examined considerably over the past 30 years. 81 The Greek periplus genre, probably beginning with the Periplus of Scylax, comprised works describing coastal sailing voyages. 82 According to Herodotus, Scylax of Caryanda sailed down the Indus and back to the Isthmus of Suez at the behest of Darius I; Hecataeus quoted fragments of a work he wrote, but it is not otherwise extant. 83 A late antique periplus of the Black Sea provides a useful insight into the variety of approaches that such writers could take. The anonymous Periplus Ponti Euxini is a compilation mainly of three works: an edition of the Periplus of Menippus of Pergamon, Arrian's Periplus Ponti Euxini, and an anonymous periegesis addressed to King Nicomedes (Ps-Scymnus, ad Nicomedem regem). 84 Menippus’ Periplus was a bare listing of stations around the coast and distances between them. 85 Arrian’s periplus supplemented the basic framework of the coastal voyage with more detailed information about the places mentioned. 86 Ps-Scymnus gave a verse description without distances, mentioning only the best known stations, but including descriptions of the hinterland as well as historical and descriptive information. 87 Not only were periploi and travellers' accounts a significant body of literature in their own right, but they were an important method of structuring particular portions of narratives within larger works. 81 See especially , Janni (1984); supported by Brodersen (1995); Whittaker (2002). 82 For example the Periplus of Hanno, Periplus Maris Erythraei, and Periplus of Scylax; Dueck (2012) 6–7. 83 Hdt. 4.44. A 4 th C BCE periplus that bears his name may be a compilation of various accounts of explorations, Shipley (2011). 84 Diller (1952) 102–46. 85 Menippus' Periplus was probably written in the late first century BCE, Diller (1952) 147–8. 86 On the structure of Arrian’s Periplus, see Rood (2011) 139–42. 87 Skymnos of Chios wrote a lost geographical survey in the first half of the second century BCE, this poetic treatment was probably composed in the early first century BCE. 30/448 In the Roman context, the linear experience of space is usually discussed in terms of travel and the itinerary genre. A number of volumes of collected papers discuss the practicalities and experience of travel in the Roman Empire. 88 The studies contained in those volumes consider evidence of travel and movement from a wide range of sources, but the most direct insight into Roman concepts of “travel directions” are the various itineraries which survive. The most detailed itinerary relevant to the Mesopotamian borderland is the Parthian Stations by Isodore of Charax. 89 This is a Greek text of the late first century describing the route from Zeugma to Alexandroplis in Arachosia, of which the most detailed section describes the route between Zeugma and Seleucia-on-the-Tigris. 90 One of the most important itineraries for study of empire-wide routes is the Itinerarium Antonini Augusti (the Antonine Intinerary). 91 This document seems to be a compilation of smaller itineraries compiled at some point in the late third century . 92 Finally , the Tabula Peutingeriana was an illustrated itinerary (itineraria picta) resembling a distorted map. 93 The extant version was made in the thirteenth century , but seems to have been based on Imperial itineraries revised several times, including at last c. 300 CE. 94 Travel based on itineraries is still a major form of spatial engagement in the 21 st century . The relationship between itineraries and two-dimensional conceptions of space can be easily imagined in a modern context by considering the linear nature of the directions given by an online mapping and navigation service such as Google Maps. Two-dimensional maps and linear directions for travel are not 88 Adams and Laurence (2001); Parker (2002) 80–84; Talbert and Brodersen (2004); Parker (2006); Adams and Roy (2007). 89 Schoff (1989); Millar (1998) 120–21. 90 Isidore 1. 91 Cuntz (1929). A second volume edited by Schnetz (1940) covers the Ravenna Cosmography and the Geographia of Guido of Pisa. 92 Perhaps around 280-90 CE, Dilke (1987) 235–6; Salway (2001) 39–43. 93 For itineraria picta, see Brodersen (2001). But note Salway’s doubts as to whether the Tabula Peutingeriana counts as such, (2001) 31–2. A digitised version of the Tabula is available online at http://www .cambridge.org/us/talbert/index.html [accessed 13 April 2014]. 94 Talbert argues that the map reflects a Diocletianic worldview: Talbert (2010) 133–6; The older view was for a fourth century date: Dilke (1987) 238. There is much debate over the status of the Peutinger Table, against Dilke's rather uncomplicated view of the Table as a map in the modern sense, see Brodersen (2001); Talbert (2010). 31/448 incompatible technologies. Different modes of conceiving of travel, routes and directions apply in different situations depending on the depth of the traveller's knowledge of the route and the area. To some degree, the physical nature of a written text encourages a list-like structure, but it is possible for an author to transcend the physical restrictions of the medium to give a sense of space through descriptive use of spatial and topological language. The existence of both two-dimensional topologies and linear segments of periploi and itineraries can be seen within the broader narrative structures of Strabo and Pliny , as will be shown in Chapter 1. 95 The Main Sources Having made these preliminary remarks, I will introduce the five Roman imperial geographical writers which will form much of the focus of my project: Strabo, Pliny the Elder, Claudius Ptolemy , the Expositio Totius Mundi et Gentium and Ammianus Marcellinus. These five sources span the chronological period of my work and include examples of scientific and descriptive geography , a variety of scales and degrees of selectivity , and both planar and linear conceptions of space. Most importantly , they all describe the Mesopotamian borderland in ways that illuminate that space and its place in the Roman imagination. Strabo Strabo was born into an elite Pontic family in Amasia around the middle of the first century BCE, just after the Roman conquest. 96 He received a Hellenistic education in Asia Minor and Rome and circulated among the intellectual elite of the Roman empire. 97 Although his major work was probably a 47 book 95 See also Dueck (2000) 40–43; Nicolet (1991) 173; Romer (1998) 21 96 Pothecary , (1997) 245, argues (against the traditional birth date of 63/64 BCE proposed by Neise in the 19 th century) that Strabo’s birth date cannot be narrowed beyond in a window from 65-50 BCE. For more detail on Strabo’s biography , see Dueck (2000) 1–30; Clarke (1997); Drijvers (1998) 279; Diller (1975). 97 Dueck (2000) 9–12; Clarke (1997) 99–101, 108; Nicolet (1991) 73. Strabo had close connections with Rome (including visits 6.2.6, 8.6.23) and may have become a Roman citizen: Dueck (2000) 7, 87–88, 96, 130–144; Pothecary (1999). 32/448 continuation of Polybius’ history , now lost except for a handful of fragments, 98 he is best known for his 17 book Geography written late in his life (around 18-24 CE) from notes compiled over his lifetime. 99 This was a work of descriptive rather than scientific geography: he intended to provide geographic and ethnographic descriptions of space with practical use, rather than scientific measurement and discussion, and mathematical calculation of space, as did many of his predecessors. 100 The Geography describes the world as it was known to contemporary Roman thought (including chapters on many areas outside the Roman world, even those as far distant as India and Scandanavia), including one of the earliest and most detailed surviving descriptions of the geography of the Mesopotamian borderland. The organisational principle of his work was the Roman imperial ideology of universal rule; the world as organised by Roman power. 101 Pliny the Elder Gaius Plinius Secundus (23-79 CE) came from a North Italian equestrian family , served several terms in the army , and held at least one provincial procuratorship, probably rising to prominence under the patronage of Vespasian and his advisor Licinius Mucianus. 102 He wrote several works over the course of his life; the last was the 37 book Natural History (Historia Naturalis), written under the Flavians and finished around 77 CE. 103 The Natural History was constructed as a catalogue of the world arrayed before the future emperor Titus and included a geographical account of the world (books 3-6). Pliny , like Strabo, is an author who is more often “consulted” than “read”. 104 Pliny’s nephew called the the Historia Naturalis an broad and 98 FGrH 9 1; Clarke (1997) 93. 99 The dating of his Geography is based on a specific reference to a time of writing (33 years after the surrender of the Alpine tribes to Tiberius in 15 BCE: 4.6.9) and the latest dateable event (the death of Juba, king of Mauretania, in 23 CE: 17.3.7). Furthermore, Dueck argues that his style of composition (well organised and with frequent cross-references) indicates a relatively short span of writing: Dueck (1999); Dueck (2000) 146– 150; Pothecary (2002). For earlier scholarship on this issue, see Drijvers (1998) 279, n.3. An example is Syme (1995); Lindsay (1997). 100 See above, p.25. Dueck (2000) 53–62; Dilke (1985) 62–65. 101 Strabo was interested in the relationship between geography and state power: 1.1.17-18. Nicolet (1991); Ando (2000) 323–29. 102 For Pliny's biography , see Beagon (2005) 1–5; Beagon (1992) 1–6; Reynolds (1986); Syme (1969). 103 For a brief summary of Pliny and the geographical aspects of his work, see Dilke (1985) 66–71. 104 To borrow the phrasing of Gibson and Morello's editorial preface, (2011) vii. 33/448 learned work, no less varied than nature itself (“opus diffusum eriditum, nec minus varium quam ipsa natura”) 105 Accordingly , the scope of his work has led to many divergent scholarly threads, especially on Roman scientific thought and his encyclopaedic project. 106 Ptolemy Ptolemy seems to have been born around the beginning of the second century CE and began his scientific career in the mid-120s in or near Alexandria. 107 He probably lived until the last quarter of the second century . Even though he lived in an intellectual centre of the Roman world at the height of the Principate, Ptolemy was also part of a scientific tradition which was firmly linked to its Hellenistic predecessors. 108 His career began with astronomical research, the Almagest, a mathematical treatise on motion of heavenly bodies. Among his works was the Geographike Hyphegesis (Manual of Geography, or usually simply Geography) a comprehensive list of coordinate data for cities, towns, rivers, mountain ranges and other such geographical features by which one could draw a map. The main body of the text was preceded by a theoretical chapter in which he identified and criticised his main source, Marinus of Tyre, and discussed the construction of an appropriate map projection to display the data he provided. Expositio The Expositio Totius Mundi et Gentium (often shortened to Expositio Totius Mundi or simply Expositio) is an anonymous text from the fourth century . 109 It was originally written in Greek, but survives 105 Pliny Ep 3.5.6. 106 For example the collections of French and Greenaway (1986) and Gibson (2011). See also Doody (2010) on the reception of his encyclopaedism, Murphy (2004) on the role of geography and ethnography in his imperialistic project, and Beagon (1992) 159–201 on his attitude towards agriculture and trade. 107 Berggren (2000) 17–20; Jones (2008). 108 Bagrow (1945) 331. 109 Rougé (1966) 9–26 established the terminus a quo as 347 CE and dated the text to 359/60 CE. An ante quem date can be determined from only non-economic descriptors of the Mesopotamian section. Exposito 22 clearly marks Mesopotamia as Roman territory (“Post hos nostra terra est”) and includes Nisibis and Amida as two of its important cities, cities which resist Persian attacks (“habentes moenia inclita, bello semper virtutem Persarum dissolvunt”). Following Julian’s disastrous campaign against Ctesiphon, Jovian handed Nisibis over the the Sasanids in the peace of 363 CE; the Expositio must antedate this concession. Furthermore, if we accept the 34/448 now as two Latin copies, the Expositio Totius Mundi et Gentium by which it is now known, and a slightly less complete copy called the Descriptio Totius Mundi. 110 It describes the Roman world in a provincial structure and some of the non-Roman world as a list of adjacent tribes and regions with varying levels of description. 111 Ammianus Marcellinus The historian Ammianus Marcellinus provides an valuable eyewitness account of the warfare in Mesopotamia during the fourth-century (specifically 354-378 CE). 112 Ammianus included several geographical digressions in his work, including on the lands and ethnography of the Saracens, on the eastern provinces, and on Persia. 113 As well as these descriptions, Sundwall argues that Ammianus is “the best source” for an investigation into the oral and written traditions of Roman mental cartography . 114 According to Sundwall, that tradition was primarily literary rather than scientific and Ammianus’ understanding of relative topology was no weaker for it. 115 4. Discussing the Chapters The dissertation is divided into three parts: Part I (Chapters 1 and 2) examines how the geographical writers understood and described the Mesopotamian borderland and the territorial, social, and political units within it. Part II (Chapters 3 and 4) shows how the Roman imperial geographic writers took Hellenistic sources of knowledge and fashioned them into a Roman geography of this inter-imperial space. Part III (Chapter 5) examines the role that physical mobility played in the geographical construction of the emendation of the doubled Edessa to Amida (see Chapter 4), the text must antedate the fall of the city to the Sasanians in 359/360. W oodman (1964) 61 dated the text slightly earlier (350-53). From Expositio on Noricum: Sinko (1904) 566, n.479. 110 Riese (1878) 104–126; Rougé (1966). 111 The non-Roman areas: Expositio 8-20. 112 Thompson (1947) 1–19; Rowell (1964) 15–29; Matthews (1989) 8–32. 113 Saracens: Amm. Marc. 14.4. Eastern provinces: Amm. Marc. 14.8. Persia: Amm. Marc. 23.6. 114 Sundwall (1996) 621–2. 115 Sundwall (1996) 639. 35/448 Mesopotamian borderland. Part I: Constructing Space Chapter 1: Naming and Placing Chapter 1 begins by examining the various ways in which the Roman Imperial geographic writers construct, delimit and denominate the physical space of the Mesopotamian borderland. I focus on the deliminations and denominations expressed by the five Roman imperial geographical writers introduced above. Strabo’s detailed descriptive geography of the land inhabited by “the Assyrians” included all of the Mesopotamian borderland, as well as the adjacent areas of Roman and Parthian space. Pliny describes the Mesopotamian borderland in two overlapping sections focusing on Roman and Parthian space respectively . Ptolemy’s mathematical geography represents the world primarily through scientific calculation, including precisely defined sections covering the space of the Mesopotamian borderland in a geographical manner. The Expositio Totius Mundi divides the borderland according to administrative structures: Roman provincial organisation and Persian dominion. Finally , the surviving books of Ammianus Marcellinus imperial history include several dedicated geographic descriptions, as well as many geographical references and asides which situate the Mesopotamian events of his historical narrative securely in a region very familiar to him. This chapter comprises an in depth examination of the way these five works represent the physical space of the Mesopotamian borderland. Chapter 2: States in the Borderland Chapter 2 zooms in on several major political entities of the Mesopotamian borderland to examine their treatment by the Roman imperial geographical writers. The representation of borderland kingdoms (Commagene and Osrhoene) and cities (Palmyra and Nisibis) and their relation to the imperial powers in these geographical works illustrate how geographical texts undercut and problematise relationships of 36/448 political allegiance and demonstrates the ambiguity in political relationships that permeates narratives of the border region. Part II: The Use of Knowledge Chapter 3: Hellenistic Knowledge Chapter 3 analyses the Hellenistic basis for the early imperial descriptions of the borderland. Strabo’s Geography , Pliny’s Natural History , and Ptolemy’s Geography are our three most complete sources for the Mesopotamian borderland. These three works drew on the tradition of Hellenistic geographic knowledge for the raw geographical material they used to describe northern Mesopotamia. As much of this knowledge was derived from the military campaigns of Alexander and the Seleucid kings, as well as from Hellenistic administrative and commercial travel the representation constructed by Roman imperial geographic writers was predicated on mobility . These writers created a representation of Mesopotamia that was delimited and denominated using the tools of Hellenistic geographic science (such as spatial measurement and systems of categorisation and naming) and that was populated with geographic and ethnographic features derived from Hellenistic movement and migration. Chapter 4: Roman Geography Chapter 4 demonstrates how the Roman imperial geographic writers used the legacy of Hellenistic knowledge to create a geography of the region centred on Roman power. By their selection and presentation of material, use of historical comparanda, and treatment of imperial power and contemporary political relationships. The negotiation between Roman and Parthian power lies at the heart of the presentation of the Mesopotamian borderland. Strabo establishes links between his Roman world and that of the fertile crescent by references to Alexander the Great, the Macedonian colonies of the Seleucid Kingdom, and contemporary political circumstances, namely the presence of the Parthian empire. Pliny constructs his narrative to portray 37/448 Provincia Syria (broadly conceived) as an ordered Roman space and the bulk of Mesopotamia as a chaotic Parthian space. In doing so, these authors appealed to both Roman political ideology and Hellenistic intellectual authority in their representation of the Mesopotamian borderland during the Principate. Part III: The Mobile Landscape Chapter 5: Commercial Factors Chapter 5 shows that the explicit or implicit depiction of commercial and economic themes construct a vision of Mesopotamia as a space of transit in networks of long distance trade. Those networks were represented through three sets of routes across the borderland. Strabo’s narrative of commercial movement on the Euphrates between Syria and Babylon emphasises inter-imperial contact and the activity of mobile pastoralists. The nomadic Skenitae were the ultimate representation of borderland movement and ambiguity: a mobile people of uncertain allegiance whose very means of subsistence vacillated between productive pastoralism and parasitic banditry . Pliny’s brief but illuminating picture of Palmyra and its connections between Babylonia and the Mediterranean highlights the centres of contact which participated in their desert routes. The Expositio Totius Mundi’s representation of the fourth-century provinces of Mesopotamia and Osrhoena astride the northern route near the Taurus shows the preeminance of that route in the late empire and particularly the importance that Nisibis attained as a focal point on inter-imperial activity . In all these cases, the representation of cross-border connections highlight the Mesopotamian borderland’s role as a space of inter-imperial movement. Chapter 6: Conclusion I conclude by bringing these factors together to show how the geographic writers place the Mesopotamian borderland in a “global” context through relations between the universal and the local. Peripheral contact zones produce cultural changes which propagate to central regions. The Mesopotamian 38/448 borderland is represented as a frontier at the edge of a Roman world and as a globalising link between the Roman and Iranian worlds. The construction of the Mesopotamian borderland in the geographical works of the Roman Empire illuminates both the relationship between Rome and the Iranian empires of the Arsacids and Sasanids and the Romans’ understanding of their own place in a global world. 39/448 Chapter 1: Placing and Naming The first task of my dissertation proper is to systematically examine our main sources of geographical information on the Mesopotamian borderland to establish in each case how these Roman imperial geographical writers named, defined and described the geographical area of the borderland. The five works discussed in this chapter (Strabo’s Geography , Pliny the Elder’s Natural History , Ptolemy’s Geography , the Expositio Totius Mundi et Gentium and Ammianus Marcellinus’s History) each discuss the space between and around the Tigris and Euphrates rivers at the limit of the Roman Empire in different ways and with different degrees of specificity . Strabo includes the Mesopotamian borderland within the land of the Assyrians and defines a geographical space of Mesopotamia between the two rivers. Pliny divides the Mesopotamian borderland between Roman and Parthian space. Ptolemy divides the space between strictly defined geographical areas using the rivers as boundaries. The Expositio Totius Mundi and Ammianus Marcellinus refer to Roman administrative regions. From this survey , two things should be clear: a) that the Roman imperial geographical writers delimited space as aggregated units according to different criteria in service of their own analytic criteria, and b) that the term “Mesopotamian borderland” is my own analytic category that does not conform precisely to the categories of any of the geographical writers under discussion. Rather, it allows the descriptions of my sources to be compared fruitfully . This does not mean that these geographical writers did not conceive of the space as a borderland, that is, a zone in which multiple boundaries and sets of boundaries overlapped and interacted in a variety of complex ways. 116 By examining the nature of way the geographical writers delimit and denominate space, we can see how they imagined the space, both geographically and otherwise (historically , culturally , militarily , economically and so on). 116 See Introduction. 40/448 Strabo All of Strabo’s description of the Mesopotamian borderland falls within the area he labels as οἱ Ἀσσύριοι, My analysis of Strabo’s conception of the borderland begins with a broad discussion of οἱ Ἀσσύριοι, then narrows to consider the relevant spaces with it: “Syria”, including Commagene and Cyrrhestice, and “Mesopotamia”, including Mygdonia, Gordyene, and the desert regions. This structure mirrors Strabo’s own, in which he first describes the broad area of οἱ Ἀσσύριοι in toto, then works through the sub-regions into which he divides that space. Throughout this discussion, I will show the fundamentally geographical basis of Strabo’s spatial divisions and the Mediterranean lens through which he views the borderland. Οἱ Ἀσσύριοι In book 16 of his Geography , Strabo describes the Arabian peninsula and the fertile crescent. He gives this latter area the name οἱ Ἀσσύριοι, thus using an ethnic term for the people of the area as a substantive for the space itself. Strabo’s οἱ Ἀσσύριοι encompassed an area from the Persian gulf, up the basin of the Tigris and Euphrates as far as the Taurus range, then west up to Phoenicia and Judaea, and included the Mesopotamian borderland (see Map 18: Ptolemy's Syrian Districts). 117 Strabo begins by defining οἱ Ἀσσύριοι according to the people and places which lay within it: Τῇ δὲ Περσίδι καὶ τῇ Σουσιανῇ συνάπτουσιν οἱ Ἀσσύριοι· καλοῦσι δ’ οὕτω τὴν Βαβυλωνίαν καὶ πολλὴν τῆς κύκλῳ γῆς, ἧς ἐν μέρει καὶ ἡ Ἀτουρία ἐστίν, ἐν ᾗπερ ἡ Νίνος καὶ ἡ Ἀπολλωνιᾶτις καὶ Ἐλυμαῖοι καὶ Παραιτάκαι καὶ ἡ περὶ τὸ Ζάγρον ὄρος Χαλωνῖτις καὶ τὰ περὶ τὴν Νίνον πεδία, Δολομηνή τε καὶ Καλαχηνὴ καὶ Χαζήνη καὶ Ἀδιαβηνή, καὶ τὰ τῆς Μεσοποταμίας ἔθνη τὰ περὶ Γ ορδυαίους καὶ τοὺς περὶ Νίσιβιν Μυγδόνας μέχρι τοῦ Ζεύγματος τοῦ κατὰ τὸν Εὐφράτην καὶ τῆς πέραν τοῦ Εὐφράτου πολλὴ ἣν Ἄραβες κατέχουσι, καὶ οἱ ἰδίως ὑπὸ τῶν νῦν λεγόμενοι Σύροι μέχρι Κιλίκων καὶ Φοινίκων καὶ Ἰουδαίων καὶ τῆς θαλάττης τῆς κατὰ τὸ Αἰγύπτιον πέλαγος καὶ τὸν Ἰσσικὸν κόλπον. 118 117 The sections of Strabo pertaining to the Mesopotamian borderland are: Strabo 16.1.21-28 (Mesopotamia itself); 16.2.3 (Commagene); 16.2.8 (Cyrrhestice). 118 Strabo 16.1.1: “The Assyrians lie next to Persis and Susiana. They call Babylonia and much of the country all round by this name, which in part is Atouria, in which is Ninus, and Apolloniatis, the Elymaei, the Paraetacae, the Chalonitis near Mt. Zagrus, the plains around Ninus, and also Dolomene and Calachene and Chazene and 41/448 By defining the space primarily according to its occupants, Strabo shows his conception of geography as historically , culturally and demographically contingent. This is particularly clear in his description of the boundaries of Syria where he shows his awareness that the spaces he describes can and are defined differently in different times and places and by different authors. 119 By defining the spaces according to the people, Strabo allows for the fluidity and movement of the populations he describes. The Greek name οἱ Ἀσσύριοι Adiabene, and the tribes of Mesopotamia near the Gordyaeans, and the Mygdonians around Nisibis as far as Zeugma on the Euphrates, and much space on the far side of the Euphrates, which is occupied by Arabians, and those people who are now called Syrians, in a particular sense, who extend as far as the Cilicians and the Phoenicians and the Judaeans and the sea that is opposite the Aegyptian Sea and the Gulf of Issus.” All translations are my own unless otherwise noted. 119 Strabo 16.2.2, quoted below , p.45. 42/448 Map 2: Strabo's Assyria ultimately derives from the Akkadian “Aššur”, the eponymous chief god of the Assyrians. Part of this area Strabo calls Ἀτουρία. Although Strabo does not make the connection, this is a Greek transliteration of Athura, the Aramaic word for the heartland of the former Assyrian empire. 120 Strictly , the Assyrian heartland comprised land adjacent to the Tigris near the Taurus, but after the Neo-Babylonian and Median conquest of Assyria in the 610s BCE, the remains of the Assyrian state apparatus which governed from Harran in north-western Mesopotamia (Roman Carrhae, now in south-east Turkey) seems to have transferred the name Athura to the entire area of northern Mesopotamia. 121 Strabo’s use of the term to refer to such a wide area was not peculiar to him. In the first century BCE, Meleagros of Gadara referred to his home town in the Decapolis (in modern Jordan) as “an Attic town among Assyrians” (Ἀτθις εν Ἀσσυριοις). 122 As we shall see in relation to Trajan's conquests in the region, the name Assyria had a powerful hold in the Greco- Roman historical imagination, but seldom a very precise geographical definition. 123 Strabo's opening description of οἱ Ἀσσύριοι is spatially defined and delimited by a list of contents: the major cities, regions and people of the region. These people and places can be geographically divided into six groups: Babylon and the surrounding area (Babylonia); Ninus (Nineveh) and the rest of Atouria/Athura proper; the people and lands between Babylon and Susa (Apolloniatis, the Elymaei, the Paraetacae, and Chalonitis near the Zagros Mountains); Adiabene and neighbouring regions (Dolomene, Calachene, Chazene); Mesopotamia (the tribes of Mesopotamia near the Gordyaeans, the Mygdonians around Nisibis as far as Zeugma, and the Arabians on the other side of the Euphrates) and Syria. Some of these locations are well known, others are not, but except for Babylonia, Mesopotamia and Syria, they all lie on the east side of the Tigris. In this opening passage, Mesopotamia is described according to the people who 120 Radt (2002) 8.249; Herzfeld (1968) 304–5. 121 Herzfeld (1968) 305. For a brief discussion of the relative spatial orientations of the Akkadians and Assyrians, see Herzfeld (1968) 57. 122 Ant. Pal. 7.417.2. 123 See the discussion of Semiramis in Chapter 4. 43/448 live there: the tribes of Mesopotamia near the Gordyaeans (τὰ τῆς Μεσοποταμίας ἔθνη τὰ περὶ Γ ορδυαίους), and the Mygdonians near Nisibis as far as Zeugma on the Euphrates (τοὺς περὶ Νίσιβιν Μυγδόνας μέχρι τοῦ Ζεύγματος τοῦ κατὰ τὸν Εὐφράτην). This approach contrasts with his later definition of the space according to geographical features, but does not conflict with it. 124 Here Strabo describes the people in the first instance (τὰ ἔθνη, τοὺς Μυγδόνας), but locates them in space with topographical markers (τῆς Μεσοποταμίας, περὶ Νίσιβιν ... μέχρι τοῦ Ζεύγματος τοῦ κατὰ τὸν Εὐφράτην). For Strabo, the space of οἱ Ἀσσύριοι is centred on Babylonia (καλοῦσι δ’ οὕτω τὴν Βαβυλωνίαν καὶ πολλὴν τῆς κύκλῳ γῆς), but his account is not specific regarding the boundaries of that sub-division, particularly in the north where it bordered Mesopotamia. 125 Strabo reports that the land of the Babylonians (ἡ χώρα τῶν Βαβυλωνίων) is bordered on the west by the Skenitai Arabs whose land stretches as far as Adiabene and Gordyene (ἀπὸ δὲ τῆς ἑσπέρας ὑπό τε Ἀράβων τῶν σκηνιτῶν μέχρι τῆς Ἀδιαβηνῆς καὶ τῆς Γ ορδυαίας). 126 This passage places the Armenians and Medes to the north of Babylonia and the Persian Gulf to the south, so Strabo’s “western” border occupies a broad, nearly 180 degree, arc from the northwest to the southeast (see Map 18: Ptolemy's Syrian Districts). Nevertheless, the specific note that the Skenitai extend as far as Gordyene, makes clear that Strabo includes the Skenitai of southern Mesopotamia in this segment. Those Skenitai thus form the southern border of Strabo’s Mesopotamia and the Tigris defines the eastern side. There are some correspondences between Strabo’s divisions and the administrative history of the region. Strabo’s conception of the land of οἱ Ἀσσύριοι as a unitary region aligns neatly with our 124 See Strabo 16.1.21, below , p.45. 125 His focus on the canal system implies that he is considers that a defining feature of the region, Strabo 16.1.9-10. 126 Strabo 16.1.8: “περιέχεται δ᾽ ἡ χώρα τῶν Βαβυλωνίων ἀπὸ μὲν τῆς ἠοῦς ὑπό τε Σουσίων καὶ Ἐλυμαίων καὶ Παραιτακηνῶν, ἀπὸ δὲ τῆς μεσημβρίας ὑπὸ τοῦ Περσικοῦ κόλπου καὶ τῶν Χαλδαίων μέχρι Ἀράβων τῶν Μεσηνῶν, ἀπὸ δὲ τῆς ἑσπέρας ὑπό τε Ἀράβων τῶν σκηνιτῶν μέχρι τῆς Ἀδιαβηνῆς καὶ τῆς Γ ορδυαίας, ἀπὸ δὲ τῶν ἄρκτων ὑπό τε Ἀρμενίων καὶ Μήδων μέχρι τοῦ Ζάγρου καὶ τῶν περὶ αὐτὸ ἐθνῶν.” (The country of the Babylonians is surrounded on the east by the Susians and Elymaeans, and Paraetacenians and, on the south by the Persian Gulf, and the Chaldæans as far as the Arabian Meseni, and on the west by the Arabians called Skenitai as far as Adiabene and Gordyaea, and on the north by the Armenians and Medes as far as the Zagrus, and the tribes about that river.) Loeb trans. 44/448 understanding of Achaemenid Persian satrapies. Much of our information about the Achaemenids comes from inscriptions dedicated by the kings themselves and inscribed on cliffs and monumental buildings. Some of these recount the deeds of the kings (especially Darius I (r. 522-486)) in the first person and some include lists of the people and places they conquered and ruled. 127 In the Persian royal inscriptions, the territory of Strabo's οἱ Ἀσσύριοι is covered by the satrapies of Bâbiruš, Athurâ and Arabâya: namely , southern Mesopotamia as formerly ruled from Babylon, northern Mesopotamia as formerly controlled by Assyria, and the steppe regions of the middle Euphrates. 128 In fact, despite the satrapies appearing in a different order in every royal inscription that lists them, the order of Bâbiruš, Athurâ and Arabâya is almost always consistent. 129 In Herodotus’ description of the Achaemenid taxation and muster districts, he refers to the area thus: Ἀπὸ Βαβυλῶνος δὲ καὶ τῆς λοιπῆς Ἀσσυρίης χίλιά οἱ προσήιε τάλαντα ἀργυρίου καὶ παῖδες ἐκτομίαι πεντακόσιοι· νομὸς εἴνατος οὗτος. 130 This area includes the lowland areas of the fertile crescent west of Media (the 10 th district) and Susa (the 8th), south of the Taurus (Armenia was part of the 13 th district) and east of Syria (the 5 th district). 131 Although he does not directly refer to Mesopotamia, that region must have been included within the rest of Assyria (τῆς λοιπῆς Ἀσσυρίης) because the bounds of the region included Syria at the time. Babylonian documents show that there was a Satrap of “Babylonia and Ebir-Nari” from the time of Cyrus. 132 Ebir-Nari (“across the river”) 127 Especially DB, DNa, DPe, DSaa, XPh; Lecoq (1997) 130–52. References to Achaemenid royal inscriptions conventionally take the form of an alphanumeric code comprising a letter denoting the king who nominally wrote it, an abbreviation denoting the location at which it was found, and a letter to differentiate between multiple inscriptions by the same ruler at the site. Thus Darius’ Behistun inscription is DB, while DPe is an inscription of Darius from a terrace at Persepolis. 128 DB col.1, lines 14-15; DPe lines 10-11; DSe lines 25-26; DNa lines 26-27; DSaa; XPh lines 22 & 25; Briant (2002) 172–3; Herzfeld (1968) 292, 357–9. Lecoq (1997) 130. 129 The exception is XPh in which Bâbiruš and Athurâ appear on line 22 and Arabâya on line 25, separated by several provinces. 130 Hdt 3.92: “From the Babylonians and the rest of Assyria, 1000 talents of silver and 500 castrated boys; this is the ninth district.” 131 Hdt 3.91 (5 th -8 th ); 3.92 (9 th -12 th ); 3.93 (13 th -16 th ). 132 Stolper (1989); Heltzer (1992); Briant (2002) 49, 884–5. “More than twenty texts dated between 535 and 525 give the title to a man with the Iranian name Gūbaru.” Stolper (1989) 289–90. 45/448 is an earlier term for the region to the west of the Euphrates (thus is often rendered Trans-Euphrates) which was still in use when Darius commissioned his palace inscriptions at Susa. Ebir-Nari comprised Syria, Phoenecia, and Cyprus, and was Herodotus’ fifth district. 133 The satrapy of Babylonia and Ebir-Nari survived the administrative changes of Darius I (522-486 BCE). 134 Between the reign of Xerxes (485-465 BCE) and the end of the Achaemenid empire, Ebir-Nari was detached from Babylonia as a separate satrapy , although it may have remained subordinate for some time. 135 Briant thinks it likely that Arbelitis, the hinterland of Arbela to the east of the Tigris (see Map 1.1), was administered as part of Athurâ during the Achaemenid period. 136 Herodotus' list is topographically vague but the Ionian historian seems to have had some degree of access to reliable information (including material derived from the text of the Behistun inscription of Darius I); his characterisation of Babylonia, Mesopotamia and Syria seems quite accurate. 137 In this conception of the region, the tradition of Greek knowledge of the space seems to have been reliably derived from Achaemenid Persian administrative units. Before Alexander’s conquest of that empire, the Greeks knew of Mesopotamia as an Achaemenid province. The Achaemenid satrapal system was inherited by the conquering Macedonians. Accounts of 133 Hdt 3.91: “Ἀπὸ δὲ Ποσιδηίου πόλιος, τὴν Ἀμφίλοχος ὁ Ἀμφιάρεω οἴκισε ἐπ’ οὔροισι τοῖσι Κιλίκων τε καὶ Συρίων, ἀρξάμενος ἀπὸ ταύτης μέχρι Αἰγύπτου, πλὴν μοίρης τῆς Ἀραβίων (ταῦτα γὰρ ἦν ἀτελέα), πεντήκοντα καὶ τριηκόσια τάλαντα φόρος ἦν· ἔστι δὲ ἐν τῷ νομῷ τούτῳ Φοινίκη τε πᾶσα καὶ Συρίη ἡ Παλαιστίνη καλεομένη καὶ Κύπρος· νομὸς πέμπτος οὗτος.” (The fifth province was the country (except for the part belonging to the Arabians, which paid no tribute) between Posideion, a city founded on the Cilician and Syrian border by Amphilochus son of Amphiaraus, and Egypt; this paid three hundred and fifty talents; in all this province was all Phoenice and the part of Syria called Palestine, and Cyprus). Loeb trans. Herodotus may have been mistaken in including Cyprus in this district. Herzfeld has the island as part of a satrapy with Cilicia, Herzfeld (1968) 309. Briant retains Herodotus’ division for the period of Darius I, Briant (2002) 488–9 (Cyprus), 498–500 (Cilicia). 134 Briant (2002) 392–3, 484, 487–90, 951–2; Herzfeld (1968) 60–1, 305. 135 The exact date of the division is unknown. “It was probably accomplished before 420, when cuneiform texts give the provincial governor of Babylonia, Gūbaru, the title “governor of the land of Akkad,” with no mention of Syria”, Stolper (1989) 297–8; Briant (2002) 578; earlier subordinate status, 487, 601. The missions of Erza and Nehemiah from Babylon to Judah attest to aspects of the fifth century administrative apparatus of Ebir-Nari, Briant (2002) 583–5. 136 Briant (2002) 719. 137 Briant (2002) 390–4, discusses the difficulties with Herodotus' list. 46/448 Alexander’s successors reveal something of the Macedonian satrapal system, but nothing of the specific geographical boundaries of the places they describe. When Perdiccas assigned the satrapies following Alexander's death in 323, he divided Babylonia and Mesopotamia and retained the division of Syria from both. 138 When the Macedonian general Antipater reassigned the satrapies in 320, Mesopotamia and Arbelitis were assigned to the same man, and remained separate from Babylonia. 139 The apparatus of Seleucid government in Babylonia and Mesopotamia during the third and second centuries BCE showed significant continuity with Persian administration, but the general trend of increasing administrative division continued. 140 Under the Seleucids, the region covered by Strabo’s οἱ Ἀσσύριοι was divided between the satrapies of Mesopotamia, Babylonia, the Districts of the Persian Gulf (Mesene and Characene), Apolloniatis, and Parapotamia. 141 These changes seem to comprise the redrawing of divisions internal to the Bâbiruš- Athurâ-Arabâya satrapal grouping but our knowledge of their boundaries is mostly implicit from their names. Early Roman imperial authors like Strabo and Pliny used their Hellenistic sources extensively , but while their testimony is often used to discuss these Macedonian satrapies, their later date and evident reworking of material makes that method problematic. Nevertheless, from the satrapal names reported in the sources, it appears that considering the plains south of the Taurus and east of the Euphrates as a broad region was administratively convenient. The mountains to the north and east (the Taurus and Zagros respectively) 138 Diodorus lists the satrapies in the following order (18.3.1-3): Egypt; Syria; Cilicia; Media (Greater); Paphlagonia and Cappadocia (and bordering lands); Pamphylia, Lycia and Greater Phyrgia; Caria; Lydia; Hellespontine Phrygia; Thrace and neighbouring tribes; Macedonia and neigbouring peoples; the kingdoms of Taxiles and Porus in India; the satrapy next to Taxilies'; the Hindu Kush (that of the Paropanisidae); Arachosia and Cedrosia; Aria and Drangine; Bactria and Sogdia; Parthia and Hyrcania; Persia; Carmania; Media (Lesser or Atropatene); Babylonia; Mesopotamia. Dexippus (FGrH, 100.8) gives much the same list, with the interesting note that he defines Babylonian and Mesopotamia thus: “Βαβυλωνίων δὲ καὶ τῆς μέσης τῶν ποταμῶν Τίγρητος καὶ Εὐφράτου τῶν μὲν Σέλευκος, τῆς δὲ Μεσοποταμίας ᾽Αρχέλαος ἦρχε.” Arrian (FGrH, 156.1.5-8) only mentions Syria. Quintus Curtius (10.10.1-6) gives the same list, but refers to the satrapies east and south of Media in general terms and omits mention of Mesopotamia or Babylonia. 139 On this occasion, Diodorus lists the satrapies as follows (18.39.5-6): Egypt; Syria; Cilicia; Mesopotamia and Arbelitis; Babylonia; Susiana; Persia; Carmania; Media; Parthia; Aria and Drangene; Bacrtria and Sogdia; Paropanisadae; the part of India bordering Paropanisadae; the kingdoms of Taxilis and Porus; Cappadocia; Greater Phyrgia and Lycia; Caria; Lydia; Hellespontine Phrygia. 140 Sherwin-White and Kuhrt (1993) 42–3, 46. 141 Sherwin-White (1993) 44; Edwell (2008) 68–9. 47/448 presented sufficiently different concerns that the conceptual separation between highlands and lowlands was maintained from at least the time of Achaemenid satrapal organisation down to Strabo’s description of the region in the Roman period. Strabo’s overall conception of οἱ Ἀσσύριοι is noteworthy in several respects. First, while the divisions of the Geography align with Persian and Seleucid administrative conceptions on a broad scale, Strabo does not seem to take the Achaemenid or Seleucid administrative structure as the basis for his sub-divisions or his narrative. The caveat here is that we are poorly informed about Seleucid administrative geography , which may have resembled Strabo’s structure to a greater degree than the extant evidence suggests. Rather, Strabo divided the space of οἱ Ἀσσύριοι on the basis of geography and ethnography , as is clear from the terms he uses in the opening section, and in those he uses in what follows. Second, on the basis of the space allocated to discussion of Babylonia alone (16.1.5-16), Strabo or his sources consider that area to be the most important part of the land of the Assyrians. 142 Strabo also devotes a lot of attention to Susiana and the tribes of the southern Zagros mountains (16.1.17-18). By contrast his description of the plains between the Tigris and the Zagros Mountains is scant. He lists several places in those plains, but he returns to few of them in detail. 143 In fact, of the two sections discussing Adiabene only the first (16.1.4) contains any significant degree of geographical detail; the second (16.1.19) is almost entirely devoted to the lengthy rivalry between the Medes, Armenians and Babylonians, and the recent intrusion of the Parthians into that realm. Syria In the second chapter of book sixteen, Strabo describes Syria. This section discusses Strabo’s outline of the region and its relation to the evidence which he provides of Seleucid administrative structure. The 142 See also the way he describes οἱ Ἀσσύριοι in the first instance (16.1.1): “καλοῦσι δ’ οὕτω τὴν Βαβυλωνίαν καὶ πολλὴν τῆς κύκλῳ γῆς” (They call Babylonia and much of the country all round by this name [οἱ Ἀσσύριοι]). The land of the Assyrians is described as “Babylonia” and “the surrounding lands”. 143 See Appendix 1.1 for more detail. 48/448 relationship between geographical description and administrative boundaries also appears in my consideration of Strabo’s description of Commagene which follows. Thereafter I examine three other of Strabo’s Syrian sub-divisions which pertain to the borderland: Cyrrhestice, Chalcidice and Parapotamia. Strabo begins his description of Syria by placing the region in its geographical context. 144 In Strabo’s narrative, the boundaries of Syria are Cilicia and the Amanus range in the north, the Euphrates and the Arabian Skenitai to the east, and Arabia Felix and Egypt to the south. 145 This corresponds with the bounds of the Roman province of Strabo’s day , and parts of his narrative explicitly discuss that spatial organisation. As we saw in his description of Mesopotamia, Strabo uses Zeugma as a reference point for long distance measurements: “ἀπὸ θαλάττης δ’ ἐπὶ τὸ ζεῦγμα τοῦ Εὐφράτου στάδιοί εἰσιν… οὐκ ἐλάττους τῶν [χιλίων καὶ] τετρακοσίων”. 146 Having defined the boundaries, Strabo divides up the interior. 147 The regions relevant to the Mesopotamian borderland are those along the Euphrates: Commagene in the Taurus foothills and, to the south, between the Amanus mountains and Orontes valleys and the Euphrates, Cyrrhestice and 144 This is Strabo’s usual method when beginning to discuss a new area, Dueck (2000) 167–68. 145 Strabo 16.2.1: Ἡ δὲ Συρία πρὸς ἄρκτον μὲν ἀφώρισται τῇ Κιλικίᾳ καὶ τῷ Ἀμανῷ· ἀπὸ θαλάττης δ’ ἐπὶ τὸ ζεῦγμα τοῦ Εὐφράτου στάδιοί εἰσιν οἱ τὸ λεχθὲν πλευρὸν ἀφορί ζοντες οὐκ ἐλάττους τῶν [χιλίων καὶ] τετρακοσίων· πρὸς ἕω δὲ τῷ Εὐφράτῃ καὶ τοῖς ἐντὸς τοῦ Εὐφράτου σκηνίταις Ἄραψι· πρὸς δὲ νότον τῇ εὐδαίμονι Ἀραβίᾳ καὶ τῇ Αἰγύπτῳ· πρὸς δύσιν δὲ τῷ Αἰγυπτίῳ τε καὶ Συριακῷ πελάγει μέχρι Ἰσσοῦ. (Syria is bounded on the north by Cilicia and Mt. Amanus; and the distance from the sea to the bridge of the Euphrates, which forms the boundary of that side, is not less than fourteen hundred stadia. It is bounded on the easy by the Euphrates and the Arabian Skenitai this side the Euphrates; and on the south by the Aegyptian and Syrian Seas as far as Issus.) Loeb trans. 146 An interpolation reiterating and specifying the two ends of this distance measurement has been rejected by most editors since Kramer (1844) (Meineke (1866); Radt (2002)). Jones (1917) notes the opinion of Kramer and Meineke but includes after στάδιοι εἰσιν the line “ἀπὸ τοῦ Ἰσσικοῦ κολπου μεχρι τοῦ zευγματος τοῦ κατα Κομμαγηνην” (from the Gulf of Issus to the bridge at Commagene). On the distance, see Radt (2002) 8.285. 147 Strabo 16.2.2: Μέρη δ’ αὐτῆς τίθεμεν ἀπὸ τῆς Κιλικίας ἀρξάμενοι καὶ τοῦ Ἀμανοῦ τήν τε Κομμαγηνὴν καὶ τὴν Σελευκίδα καλουμένην τῆς Συρίας, ἔπειτα τὴν κοίλην Συρίαν, τελευταίαν δ’ ἐν μὲν τῇ παραλίᾳ τὴν Φοινίκην, ἐν δὲ τῇ μεσογαίᾳ τὴν Ἰουδαίαν. ἔνιοι δὲ τὴν Συρίαν ὅλην εἴς τε Κοιλοσύρους [καὶ Σύρους] καὶ Φοίνικας διελόντες τούτοις ἀναμεμῖχθαί φασι τέτταρα ἔθνη, Ἰουδαίους Ἰδουμαίους Γ αζαίους Ἀζωτίους, γεωργικοὺς μέν, ὡς τοὺς Σύρους καὶ Κοιλοσύρους, ἐμπορικοὺς δέ, ὡς τοὺς Φοίνικας. (W e set down as parts of Syria, beginning at Cilicia and Mt. Amanus, both Commagene and the Seleucis of Syria, as the latter is called; and then Coele-Syria, and last, on the seaboard, Phoenicia, and, in the interior, Judaea. Some writers divide Syria as a whole into Coelo-Syrians and Syrians and Phoenecians, and say that four other tribes are mixed up with these, namely , Judaeans, Idumaeans, Gazaeans, and Azotians, and that they are partly farmers as the Syrians and Coelo-Syrians, and partly merchants, as the Phoenicians.) Loeb trans. Radt (2002) 8.285–86. 49/448 Chalcidike. Strabo’s description of these areas contributes to our understanding of his border between Syria and Mesopotamia as well as our understanding of the political dynamics of the borderland. In early years of the Roman province, imperial control was focused on the coast, while much of the interior regions were indirectly controlled by client kings. The kingdom of Commagene and various small political units within Cyrrhestice and Chalcidike are better considered as part of the borderland in the first century CE. 148 Map 1.4 shows the sites named by Strabo in his description of Commagene, Cyrrhestice, Chalcidice and Parapotamia. 148 The case of Commagene illustrates the political circumstances that obtained for small kingdoms in the borderland between the Roman and Parthian empires. This is discussed in Chapter 2. 50/448 Map 3: Strabo’s Syria Strabo divides Syria according to Seleucid administrative organisation reported by Posidonius (c.135 - 51 BCE). The works of this native Syrian from Apamea who travelled widely in the Mediterranean were a major source for Strabo, particularly regarding places in Posidonius’ homeland, with which we might expect he was well acquainted. 149 Posidonius reports that Seleucis and Coele Syria were each divided into four satrapies: Ἡ δὲ Σελευκὶς ἀρίστη μέν ἐστι τῶν λεχθεισῶν μερίδων, καλεῖται δὲ τετράπολις καὶ ἔστι κατὰ τὰς ἐξεχούσας ἐν αὐτῇ πόλεις, ἐπεὶ πλείους γέ εἰσι, μέγισται δὲ τέτταρες, Ἀντιόχεια ἡ ἐπὶ Δάφνῃ καὶ Σελεύκεια ἡ ἐν Πιερίᾳ καὶ Ἀπάμεια δὲ καὶ Λαοδίκεια, αἵπερ καὶ ἐλέγοντο ἀλλήλων ἀδελφαὶ διὰ τὴν ὁμόνοιαν, Σελεύκου τοῦ Νικάτορος κτίσματα· ἡ μὲν οὖν μεγίστη τοῦ πατρὸς αὐτοῦ ἐπώνυμος, ἡ δ’ ἐρυμνοτάτη αὐτοῦ, αἱ δ’ ἄλλαι ἡ μὲν Ἀπάμεια τῆς γυναικὸς αὐτοῦ Ἀπάμας, ἡ δὲ Λαοδίκεια τῆς μητρός. οἰκείως δὲ τῇ τετραπόλει καὶ εἰς σατραπείας διῄρητο τέτταρας ἡ Σελευκίς, ὥς φησι Ποσειδώνιος, εἰς ὅσας καὶ ἡ κοίλη Συρία, [εἰς μίαν δ’ ἡ Μεσοποταμία.] 150 Unfortunately it is unclear from the administrative organisation Posidonius describes whether Cyrrhestice was a geographical or administrative area. Cohen summarises the debate over the possibilities: 1) Cyrrhestice was a geographical area overlapped by four satrapies corresponding to the four cities of the tetrapolis of Seleucia, one city per satrapy; 2) Cyrrhestice was a separate satrapy from the four satrapies of the tetrapolis; 3) Cyrrhestice was one of the four satrapies of Seleucis along with Chalcidice and two satrapies centred on Antioch and Apameia. 151 Considering the relatively close proximity of the four cities of the Tetrapolis, this last case seems the most sensible from an administrative perspective. 152 Regardless of the administrative reality of the Seleucid satrapal organisation in North Syria, Commagene, Cyrrhestice and Chalcidice were also 149 Inwood, “Poseidonius”, BNP . Strabo notes Poseidonius’s Apamean origin in his description of that city (16.2.10). Strabo’s sources are discussed in Chapter 3. 150 Strabo 16.2.4: “Not only is Seleucis the best of the above-mentioned parts [of Syria], but it is also called a tetrapolis, and it is, because it has outstanding cities in it. It has several, but the four largest are Antiocheia near Daphne, Seleuceia in Pieria, Apameia and Laodiceia; and these cities, all founded by Seleucus Nicator, used to be called sisters, because of their concord with one another. The largest is named after his father, most strongly fortified after himself, and the others, Apameia, after his wife Apama and Laodiceia, after his mother. As appropriate to the Tetrapolis, Seleucis was also divided into four satrapies, as Poseidonius says, Coele-Syria into the same number, [and Mesopotamia into only one]. Editors suspect this final clause, Jones (1971) 7.240 n.3; Radt (2002) 8.288 (app. crit. at 4.308). 151 Cohen (2006) 29. 152 Jones considers this the most likely , (1971) 241–2, n.21, as does Grainger (1990) 40–41. 51/448 geographical regions with a territorial extent which can be described, albeit imprecisely . Commagene was the largest client kingdom in northern Syria and lay in the foothills of the Amanus and Taurus ranges on the west bank of the Euphrates. 153 In the first century BCE and the first century CE, the political fortunes of the kingdom of Commagene fluctuated with the changing eastern policies of the Roman military dynasts and Emperors who held imperium in the Near East. 154 Strabo’s description illustrates this: καθ’ ἕκαστα δὲ ἡ Κομμαγηνὴ μικρά τίς ἐστιν· ἔχει δ’ ἐρυμνὴν πόλιν Σαμόσατα ἐν ᾗ τὸ βασίλειον ὑπῆρχε, νῦν δ’ ἐπαρχία γέγονε· χώρα δὲ περίκειται σφόδρα εὐδαίμων, ὀλίγη [δέ]. Ἐνταῦθα δὲ νῦν ἐστι τὸ ζεῦγμα τοῦ Εὐφράτου· κατὰ τοῦτο δὲ Σελεύκεια ἵδρυται φρούριον τῆς Μεσοποταμίας προσωρισμένον ὑπὸ Πομπηίου τῷ Κομμαγηνῷ, ἐν ᾧ τὴν Σελήνην ἐπικληθεῖσαν Κλεοπάτραν Τιγράνης ἀνεῖλε, καθείρξας χρόνον τινὰ ἡνίκα τῆς Συρίας ἐξέπεσεν. 155 Strabo includes two notes regarding Roman manipulation of the political geography . The first is that in Strabo’s time, the kingdom of Commagene had become a province (νῦν δ’ ἐπαρχία γέγονε). Pothecary has shown that Strabo’s use of νῦν and similar expressions is somewhat imprecise and cannot be used to accurately date the events mentioned in his Geography; 156 however, in this case, we have the testimony of Tacitus to assist us. Tacitus writes that Germanicus annexed Commagene to Syria in 18 CE after the death of the previous king, Antiochus III Eupator. 157 This was not the only time Commagene was transferred to 153 Strabo also locates Commagene in his description of the Taurus (11.12.2): ...τὸ δὲ Ἀμανὸν ὄρος μέχρι τοῦ Εὐφράτου καὶ τῆς Μελιτηνῆς πρόεισι, καθ’ ἣν ἡ Κομμαγηνὴ τῇ Καππαδοκίᾳ παράκειται.·(...whereas the mountain Amanus extends to the Euphrates River and Melitine, where Commagene lies adjacent to Cappadocia.) Loeb trans. 154 See Chapter 2. 155 Strabo 16.2.3: “In particular, Commagene is small. It has a fortified city , Samosata, in which lay the royal palace, but now it is a province. A small, but very fertile land lies around. In this place is now a bridge over the Euphrates (or Zeugma on the Euphrates). Opposite this lies Seleucia, a fortress of Mesopotamia, added to the dominion of Commagene by Pompey , in which Tigranes killed Selene, surnamed Cleopatra, who had been trapped there for some time when she was driven out of Syria.” Radt (2002) 8.286–88. 156 Pothecary (1997). 157 Annexation: Tac. Ann. 2.56 (Commagenis Q. Servaeus praeponitur, tum primum ad ius praetoris translatis.) Death of Antiochus III Eupator: Tac. Ann. 2.42 (per idem tempus Antiocho Commagenorum, Philopatore Cilicum regibus defunctis turbabantur nationes, plerisque Romanum, aliis regium imperium cupientibus.) Radt (2002) 8.286; Millar (1993) 52. 52/448 the direct control of the Roman administration, but it was the only time Strabo could have known about. 158 The second note is that Pompey had adjusted the borders of the kingdom during his eastern settlement; Seleucia was placed within the borders of Commagene by Pompey (Σελεύκεια... προσωρισμένον ὑπὸ Πομπηίου τῷ Κομμαγηνῷ). Strabo describes Seleucia as a fortress of Mesopotamia (φρούριον τῆς Μεσοποταμίας) and places it near Zeugma (or near a zeugma: κατὰ τοῦτο i.e. τὸ ζεῦγμα τοῦ Εὐφράτου). In fact, this Seleucia is probably Zeugma itself, which Pliny also knew as Seleucia on the Euphrates (Seleucia ad Euphraten). 159 Pompey’s assignment of the fortress to Commagene and Germanicus’ annexation of the kingdom indicate the reach of Roman administrative power from the directly controlled core of provincia Syria. The other Commagenean location mentioned by Strabo is Samosata on the west bank of the Euphrates river, now beneath the waters of the Atatürk dam’s reservoir. 160 This site was inhabited from pre- historic times and may have been founded as Samosata either by Samos I in the middle of the third century BCE or by Samos II in the second half of the second century BCE. 161 Strabo calls Samosata a polis and locates the royal residence there (τὸ βασίλειον); 162 Syme disputes the importance Strabo attaches to the site, arguing that Strabo confused Zeugma and Samosata. 163 However, there is ample evidence that Samosata was 158 Commagene was returned to the nominal control of its royal house by Caligula in 38 CE, only to be annexed a second time by Vespasian in 72 CE. See Chapter 2. 159 Pliny NH 5.82. Cohen (2006) 190–96. Syme, (1995) 101, introduces a letter of Cicero of 54 BCE in which the orator claims to have removed Zeugma from Commagene after Pompey’s assignment. Cic. Ad Q. Fr. 2.10.2-3: “De Commageno, quod rem totam discusseram, mirifice mihi et per se et per Pomponium blanditur Appius; videt enim, hoc genere dicendi si utar in ceteris, Februarium sterilem futurum. Eumque lusi iocose satis, neque solum illud extorsi oppidulum, quod erat positum in Euphrati Zeugmate, sed praeterea togam sum eius praetextam, quam erat adeptus Caesare consule, magno hominum risu cavillatus. 3. "Quod vult," inquam, "renovari honores eosdem, quo minus togam praetextam quotannis interpolet, decernendum nihil censeo; vos autem, homines nobiles, qui Bostrenum praetextatum non ferebatis, Commagenum feretis?" Genus vides et locum iocandi. Multa dixi in ignobilem regem, quibus totus est explosus.” For more on Zeugma, see Appendix 1.20. 160 Kennedy (1998a) 17–18. 161 Cohen (2006) 187–89; Sullivan (1977) 751–52. 162 Strabo 16.2.3, quoted above, p.45. 163 Syme (1995) 95–110. 53/448 an important city at a crossing point of the Euphrates. 164 In the sentence “ἐνταῦθα δὲ νῦν ἐστι τὸ ζεῦγμα τοῦ Εὐφράτου· κατὰ τοῦτο δὲ Σελεύκεια ἵδρυται φρούριον τῆς Μεσοποταμίας...”, Syme takes ἐνταῦθα and κατὰ τοῦτο to refer to Samosata specifically , whereas the subject of the preceding sentence is Commagene as a whole, not its capital specifically . Thus “Seleucia, the fortress of Mesopotamia” need only lie in some relation to Commagane, not to Samosata. Exactly what relation Strabo intends is ambiguous owing to the topological multivalancy of κατὰ with the accusative which can mean opposite (as Syme takes it) or downstream. The latter fits the probable identification of Strabo’s φρούριον τῆς Μεσοποταμίας with Seleucia on the Euphrates, that is, a fortress at the westernmost of the twin towns at Zeugma itself. 165 Strabo is guilty of topological vagueness in his placement of the crossing and the fortress, not of confusing Zeugma and Samosata. 166 Nevertheless Strabo’s inclusion of this fortress of Mesopotamia (the space east of the Euphrates) as a possession of the king of Commagene (in Syria, west of the Euphrates) emphasises that while the river was a geographical boundary , it need not have been a political barrier. After Commagene, Strabo describes Antiocheia before directing his attention to the bounds of Cyrrhestice. 167 To the north of Cyrrhestice lay Mount Amanus and Commagene, to the south, Chalcidice, to the west, the lands of Antioch and Apamea and to the east, the Euphrates. 168 Within this region, Strabo 164 Cohen (2006) 188; Comfort et al. (2000) 113. Josephus (BJ 7.224) knows Samosata as the main city in Commagene in 72-73 CE and locates it at a crossing on the Euphrates. It also appears as a crossing in the Antonine Itinerary (186.5), for which, see Chapter 5. 165 Kennedy (1998b) 37; Wagner (1976) esp. 40–51. 166 However, Strabo Mesopotamian fortress is problematic as noted above, p.45. 167 Strabo 16.2.7: ἀνάπλους δ’ ἐκ θαλάττης ἐστὶν εἰς τὴν Ἀντιόχειαν αὐθημερόν. πρὸς ἕω δ’ ὁ Εὐφράτης ἐστὶ καὶ ἡ Βαμβύκη καὶ ἡ Βέροια καὶ ἡ Ἡράκλεια τῇ Ἀντιοχείᾳ, πολίχνια τυραννούμενά ποτε ὑπὸ Διονυσίου τοῦ Ἡρακλέωνος. διέχει δ’ ἡ Ἡράκλεια σταδίους εἴκοσι τοῦ τῆς Ἀθηνᾶς ἱεροῦ τῆς Κυρρηστίδος. (Inland voyages from the sea to Antiocheia are made on the same day one starts. To the East of Antiocheia are the Euphrates, as also Bambyke and Beroea and Heracleia, small towns once ruled by the tyrant Dionysius, the son of Heracleon. Heracleia is twenty stadia distant from the temple of Athena Cyrrhestis.) Radt (2002) 8.291. On Cyrrhestice, see Schottky , “Kyrrhestike” NP; Frezouls ANRW II 8.164-97; Cohen (2006) 181–84; Honigmann RE “Κυρρηστική”. 168 Strabo 16.2.8: “Εἶτα ἡ Κυρρηστικὴ μέχρι τῆς Ἀντιοχίδος· ἀπὸ δὲ τῶν ἄρκτων ἐστὶ τό τε Ἀμανὸν πλησίον καὶ ἡ Κομμαγηνή· συνάπτει δὲ τούτοις ἡ Κυρρηστικὴ μέχρι δεῦρο παρατείνουσα. ἐνταῦθα δ’ ἐστὶ πόλις Γίνδαρος, ἀκρόπολις τῆς Κυρρηστικῆς καὶ λῃστήριον εὐφυές, καὶ Ἡράκλειόν τι καλούμενον πλησίον ἱερόν· περὶ οὓς τόπους ὑπὸ Ὀυεντιδίου Πάκορος διεφθάρη, ὁ πρεσβύτατος τῶν τοῦ Παρθυαίου παίδων, 54/448 placed Gindaros and a Heracleion, but he strongly implies that Bambyke, Beroea and Heracleia are to be included. Strabo describes these three small cities (πολίχνια) as lying to the east of Antiochia (πρὸς ἕω) and between that region and the Euphrates. 169 The locations of Bambyke and Beroea (modern Aleppo) are well known, but that of Heracleia is not. 170 Gindaros was probably named for the Macedonian city Genderos; the Hellenistic name is preserved in the modern town of Jindires. 171 Strabo notably omits the eponymous Macedonian foundation of Cyrrhos, the trace of which may remain in his reference to the temple of Athena Cyrrhestis, 20 stadia from Heracleia. 172 Following his description of Cyrrhestice, Strabo’s narrative again moves to the coast, this time to describe Laodiceia (16.2.9), before returning inland to treat Apamea (16.2.10) then Chalcidice: Ὅμορος δ’ ἐστὶ τῇ Ἀπαμέων πρὸς ἕω μὲν ἡ τῶν φυλάρχων Ἀράβων καλουμένη Παραποταμία καὶ ἡ Χαλκιδικὴ ἀπὸ τοῦ Μασσύου καθήκουσα καὶ πᾶσα ἡ πρὸς νότον τοῖς Ἀπαμεῦσιν, ἀνδρῶν σκηνιτῶν τὸ πλέον· παραπλήσιοι δ’ εἰσὶ τοῖς ἐν τῇ Μεσοποταμίᾳ νομάσιν· ἀεὶ δ’ οἱ πλησιαίτεροι τοῖς Σύροις ἡμερώτεροι καὶ ἧττον Ἄραβες καὶ σκηνῖται, ἡγεμονίας ἔχοντες συντεταγμένας μᾶλλον, καθάπερ ἡ Σαμψικεράμου καὶ ἡ Γαμβάρου καὶ ἡ Θέμελλα καὶ ἄλλων τοιούτων. 173 In this region Strabo focuses on the Skenitai, their political constructions and their relationship to ἐπιστρατεύσας τῇ Συρίᾳ. τῇ δὲ Γινδάρῳ συνάπτουσιν αἱ Πάγραι τῆς Ἀντιοχίδος, χωρίον ἐρυμνὸν κατὰ τὴν ὑπέρθεσιν τοῦ Ἀμανοῦ τὴν ἐκ τῶν Ἀμανίδων πυλῶν εἰς τὴν Συρίαν κείμενον. ὑποπίπτει μὲν οὖν ταῖς Πάγραις τὸ τῶν Ἀντιοχέων πεδίον...” (Then there is Cyrrhestice as far as Antiocheia. To the north is the Amanus and Commagene. Cyrrhestice, reaching that far, borders on these. Here is the city of Gindaros, the acropolis of Cyrrhestice and useful for bandits, and nearby a place called Heracleia. Around these places Pacorus, eldest child of the Parthian king, was killed by V entidius when he marched against Syria. Pagrai in Antiocheia is on the border of Gindaros, a fortified place near the pass over the Amanus between the Amanus Gates and Syria. The plain of the Antiocheians lies under Pagrai...) 169 Strabo 16.2.7, quoted above, n.129. 170 Bambyke: Appendix 1.4. Heracleia: Cohen (2006) 171–72; Radt (2002) 8.290. 171 Pliny , NH 5.81. Cohen (2006) 170–71; Jones (1971) 262, n.48; Radt (2002) 8.291; Kettenhofen, “Gindaros”, EIran. Ptolemy (5.15.15) places Gindaros in Seleuceia rather than Cyrrhestice. 172 Cyrrhos: Cohen (2006) 181–84; Frezouls (1954). For the the temple of Athena Cyrrhestis: Strabo 16.2.7 (quoted above, n.129); Radt (2002) 8.291. ‘Kyrrhos’ PECS, p.473. 173 Strabo 16.2.11: “Bordering on the land of the Apameians, to the east, is the land of Arab chieftans called Parapotamia as also Chalcidice coming down from Massyas, and all the country to the south of the Apameians, which is full of tent-dwelling (Skenitai) people. These are similar to the nomads in Mesopotamia. For those nearer to the Syrians are always more civilised, and the Arabs and Skenitai less so, having leadership that is better organised, such as Sampsiceramus, Gambarus, Themellas, and other such rulers.” Radt (2002) 8.297. 55/448 neighbouring states. 174 He omits mention of several settlements known to have existed in the region in the Hellenistic period, including Chalcis on the Belos, the numerous towns and fortresses along the Euphrates noted by Isidore of Charax, and Dura Europos, at this time a Parthian fortress city . 175 Of particular interest in this passage is the way Strabo describes Parapotamia and Chalcidice. The former is defined by the presence of Arabian political units and the Euphrates (ἡ τῶν φυλάρχων Ἀράβων καλουμένη Παραποταμία), the latter by a geographic relationship (ἡ Χαλκιδικὴ ἀπὸ τοῦ Μασσύου καθήκουσα). 176 All of these areas are explicitly east of Apamea and implicitly south of Cyrrhestice. Furthermore, Chalcidice is “down from Massyas”, that is, downhill from the Bekaa Valley to the south. 177 Strabo places Parapotamia east of Chalcidice and along the Euphrates, perhaps as far as Babylonia, but beyond that is unclear. 178 Strabo’s brief notice of the existence of Parapotamia does not alleviate the obscurity of the region. It 174 This is also the manner in which he treats the southern parts of Mesopotamia, see above, p.72. 175 Chalcis: Pliny NH 5.81; App. Syr. 57; Cohen (2006) 143–45. Euphrates towns: Isidore 1. Dura Europus: Arnaud (1986); Cohen (2006) 156–69; Edwell (2008) 68–69; Polyb. 5.48.13-16; see Appendix 1.9. 176 In fact, this region of the North Arabian steppe is divided into three parts. A third section is defined by a cardinal reference from a geographic point and by its inhabitants (πᾶσα ἡ πρὸς νότον τοῖς Ἀπαμεῦσιν, ἀνδρῶν σκηνιτῶν τὸ πλέον). There is a curious parallel between this tripartite division and that of Mesopotamia in the previous chapter of book 16. There, Strabo divides Mesopotamia proper into three parts using relative topology , physical geography and demography as references: One part is along the mountains (ἡ παρόρειος, Strabo 16.1.23), another is inhabited by the Gordyeans (τὰ τῶν Γ ορδυαίων χωρία, Strabo 16.1.24), and the last is towards the south and further from the mountains (τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα καὶ ἀπωτέρω τῶν ὀρῶν, Strabo 16.1.26). Each of these parts uses a different style of geographic division. The first, ἡ παρόρειος, uses a category of physical geography (the mountains) as a reference. The second, τὰ τῶν Γ ορδυαίων χωρία, describes the space according to the population (the Gordyeans) which inhabits it. The last, τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα καὶ ἀπωτέρω τῶν ὀρῶν, uses two referential features: relative spatial topology (cardinal directions) and physical geography (the mountains). In his description of Syria, the three areas are defined by political and population units, by geography , and by relative spatial topology and human geography. The exact categories correspond quite closely , but that correspondence is less important than the stylistic choice to divide space into three areas each defined and described according to different geographical criteria. 177 Strabo describes the Massyas Plain at 16.2.18. Massyas/the Bekaa valley is south of Apamea, just off the bottom of Map 4: The Geographical Boundaries of Strabo's Mesopotamia, above, p.59. Massyas contains a second Chalcis, under mount Libanus, possibly at modern Gerrha, Cohen (2006) 239–42; Butcher (2003) 93; Myers (2010) 82–83. Strabo describes Coele-Syria as “above Seleucis” (ὑπὲρ τῆς Σελευκίδος) at 16.2.21. 178 Strabo writes that Babylonia writes is bounded on the west by Skenitai: “Περιέχεται δ’ ἡ χώρα τῶν Βαβυλωνίων... ἀπὸ δὲ τῆς ἑσπέρας ὑπό τε Ἀράβων τῶν σκηνιτῶν μέχρι τῆς Ἀδιαβηνῆς καὶ τῆς Γ ορδυαίας...” (16.1.8). At 16.1.27, Strabo uses τὴν παραποταμίαν to refer to the land along the river which merchants often avoided due to the high tariffs imposed by the local rulers. It is unclear whether he intends to use “parapotamia” to refer to a region or administrative area, or simply as a description of the space. 56/448 was an administrative unit under the Seleucids and was maintained as such by the Parthians, probably governed from Dura Europus by both. 179 Polybius mentions its existence during the campaign of Molon in 222 BCE, who “occupied Parapotamia as far as the city of Europus and Mesopotamia as far as Dura.” 180 These lines describe activities undertaken by Molon’s forces while he was at Seleucia on the Tigris. Polybius’ “Europos” is Dura Europos, while “Dura” is a town on the Tigris. 181 Several loan documents of the Parthian period attest to the presence of officials responsible for the administrative region of Parapotamia, including a responsibility for mediating relations with local pastoralist tribes. 182 Parapotamia continued to be an administrative region into the Roman period, as attested by a fragmentary letter from the reign of Elagabalus (218-222 CE) found at Dura Europos. 183 In his description of Syria, Strabo proceeds according to his usual method of giving a general description of a space followed by a narrower, closer examination of sub-divisions within that space. This gives a coherent structure to Strabo’s work by establishing a geographical context and framework for his narrative. These broad and narrow spatial divisions need not have the same basis. Strabo’s “Syria” is a geographical region and a Roman administrative category , while its sub-divisions are defined by a mix of geographic, political, and perhaps cultural boundaries. This is also evident in the case of Mesopotamia. Mesopotamia The basis for Strabo’s delimitation of Mesopotamia is geographical. I begin by exploring the bounds of that space; three of which (east, west and south) are relatively easily determined (for these, see Map 2: Strabo's Assyria). The northern boundary of Strabo’s space is more difficult to place. I first consider two of 179 Chaumont (1984) 82, 91–92; Arnaud (1986) 137; Sherwin-White (1993) 44; Edwell (2008) 68–69. 180 Polyb. 5.48.16: “τὴν μὲν Παραποταμίαν μέχρι πόλεως Εὐρώπου κατέσχε, τὴν δὲ Μεσοποταμίαν ἕως Δούρων.” 181 Walbank (1957) 1.579–80. Dura on the Tigris: Polyb. 5.52.2; Amm. Marc. 25.6.9; Walbank (1957) 1.582. 182 Papyrii of the Parthian period which mention Parapotamia: P . Dura 18 (87 CE), P . Dura 19 (121 CE) and P . Dura 20 (121 CE). Arnaud (1986) 137–38. 183 P . Dura 55; Edwell (2008) 69. 57/448 his sub-divisions, Mygdonia and Gordyene, before returning to that boundary . Finally , I consider the southern sub-division and its pastoralist inhabitants. Strabo’s description of Mesopotamia concludes the first chapter of book 16. 184 Whereas in his initial survey of οἱ Ἀσσύριοι he describes the area by reference to the people there (such as τὰ τῆς Μεσοποταμίας ἔθνη), here he defines the region according to its boundaries, shape and dimensions. 185 Μεσοποταμία δ’ ἀπὸ τοῦ συμβεβηκότος ὠνόμασται· εἴρηται δ’ ὅτι κεῖται τοῦ Εὐφράτου μεταξὺ καὶ τοῦ Τίγριος καὶ διότι ὁ μὲν Τίγρις τὸ ἑωθινὸν αὐτῆς μόνον κλύζει πλευρόν, τὸ δ’ ἑσπέριον καὶ νότιον ὁ Εὐφράτης· πρὸς ἄρκτον δὲ ὁ Ταῦρος ὁ τοὺς Ἀρμενίους διορίζων ἀπὸ τῆς Μεσοποταμίας. τὸ μὲν οὖν μέγιστον ὃ ἀφίστανται διάστημα ἀπ’ ἀλλήλων τὸ πρὸς τοῖς ὄρεσίν ἐστι· τοῦτο δ’ ἂν εἴη τὸ αὐτὸ ὅπερ εἴρηκεν Ἐρατοσθένης, τὸ ἀπὸ Θαψάκου, καθ’ ὃ ἦν τὸ ζεῦγμα τοῦ Εὐφράτου τὸ παλαιόν, ἐπὶ τὴν τοῦ Τίγριος διάβασιν, καθ’ ἣν διέβη Ἀλέξανδρος αὐτόν, δισχιλίων τετρακοσίων· τὸ δ’ ἐλάχιστον μικρῷ πλέον τῶν διακοσίων κατὰ Σελεύκειάν που καὶ Βαβυλῶνα. διαρρεῖ δ’ ὁ Τίγρις τὴν Θωπῖτιν καλουμένην λίμνην κατὰ πλάτος μέσην· περαιωθεὶς δ’ ἐπὶ θάτερον χεῖλος κατὰ γῆς δύεται μετὰ πολλοῦ ψόφου καὶ ἀναφυσημάτων, ἐπὶ πολὺ δ’ ἐνεχθεὶς ἀφανὴς ἀνίσχει πάλιν οὐ πολὺ ἄπωθεν τῆς Γ ορδυαίας· οὕτω δὲ σφοδρῶς διεκβάλλει τὴν λίμνην, ὥς φησιν Ἐρατοσθένης, ὥστε ἁλμυρὰν αὐτὴν οὖσαν καὶ ἄνιχθυν γλυκεῖαν κατὰ τοῦτ’ εἶναι τὸ μέρος καὶ ῥοώδη καὶ ἰχθύων πλήρη. 186 This construction by which Strabo relates the name of Mesopotamia to its geography , ἀπὸ τοῦ συμβεβηκότος (“from an accidental characteristic”), is common in the Geography and signals Strabo’s participation in the tradition of Hellenistic scholarship. 187 An “accidental characteristic” is a characteristic of 184 Strabo 16.1.21-28. The implications of this will be discussed in Chapter 6. 185 Cardinal directions: Strabo 16.1.21. Size and shape: Strabo 16.1.22. 186 Strabo 16.1.21: “Mesopotamia receives its name from an accidental characteristic. As I have said, it lies between the Euphrates and the Tigris; and the Tigris washes its eastern side only , whereas the Euphrates washes its western and southern sides; and on the north is the Taurus, which separates Armenia from Mesopotamia. Now the greatest distance by which the two rivers are separated is that towards the mountains; and this distance might be the same as that stated by Eratosthenes — I mean that from Thapsacus, where was the old bridge of the Euphrates, to the crossing of the Tigris, where Alexander crossed it — two thousand four hundred stadia; but the shortest distance between the two rivers is somewhere in the neighbourhood of Seleuceia and Babylon, slightly more than two hundred stadia. The Tigris flows through the middle of Lake Thopitis, as it is called, in the direction of its breadth; and, after traversing it to the opposite shore, it sinks underground with upward blasts and a loud noise; and having flowed for a considerable distance invisible, it rises again not far away from Gordyaea; and it traverses the lake so forcefully , as Eratosthenes says, that, although the lake elsewhere is briny and without fish, yet in this part it is fresh, runs like a river, and is full of fish.” 187 It appears 18 times: Strabo 4.1.7; 5.3.6 (ἑξῆς δ’ ἐν ἑκατὸν σταδίοις τῷ Κιρκαίῳ Ταρρακῖνα ἐστί, Τραχίνη καλουμένη πρότερον ἀπὸ τοῦ συμβεβηκότος.); 6.1.12; 7a.1.30, 44; 8.3.2, 3, 24; 9.2.17, 25; 9.3.13 (Ἑξῆς γὰρ ἐν τῇ παραλίᾳ μετὰ τὴν Ἀντικύραν πολίχνιόν ἐστιν Ὀπισθομάραθος· εἶτ’ ἄκρα Φαρύγιον ἔχουσα ὕφορμον· εἶθ’ ὁ λιμὴν ὕστατος ὁ προσαγορευθεὶς μυχὸς ἀπὸ τοῦ συμβεβηκότος, ὑπὸ τῷ Ἑλικῶνι καὶ 58/448 the so-named thing that is not essential to its nature. In this philosophical sense, the presence of the two rivers does not fundamentally change the space (one could imagine the Syrian steppe without the two rivers), but it is a non-separable, or permanent, “accident” from which the space gets its name. 188 Accordingly , the Tigris (Τίγρις τὸ ἑωθινὸν αὐτῆς μόνον κλύζει πλευρόν) and Euphrates (τὸ δ’ ἑσπέριον καὶ νότιον ὁ τῇ Ἄσκρῃ κείμενος.); 9.4.5; 9.5.18; 16.1.21; 16.2.15; 16.4.6, 7, 24. For more on Strabo’s participation in the Greek intellectual tradition, see Chapter 3. 188 Porphyry , Eisagoge 5 (Barnes (2003) 220–35 discusses the passage in relation to Aristotle’s thought); Dionysius Thrax (Techne grammatike 12) calls this a “pheronym”: “φερώνυμον δέ ἐστι τὸ ἀπό τινος συμβεβηκότος τεθέν, ὡς Τισαμενός καὶ Μεγαπένθης” (A pheronymous noun is derived from some event, such as Tisamenos (having avenged), Megapenthēs (greatly sorrowing), Kemp (1986)). 59/448 Map 4: The Geographical Boundaries of Strabo's Mesopotamia Εὐφράτης) delimit Strabo’s Mesopotamia to the east and west, and their crossings provide points by which Strabo defines its widest extent. Both rivers arise in Armenia, leading Strabo to describe their courses (especially the Euphrates) in his book on that region. 189 Those descriptions remain consistent in their definition of Mesopotamia in terms of the two rivers. However, the extent to which Strabo correctly understood the source of the Tigris is unclear. Strabo discusses the source of the Tigris on three occasions. 190 In book 16 he reports that the source is near Gordyene (οὐ πολὺ ἄπωθεν τῆς Γ ορδυαίας) and in book 11 that it emerges near Chalonitis (κατὰ τὴν Χαλωνῖτιν). 191 As the only place named Chalonitis known to us is south of Adiabene, near Babylonia, this latter topological reference must be either an error or an unknown Chalonitis which does not help us locate the source of the river. 192 Until recently , the Tigris passed under the Taurus from Lake Hazar and into Mesopotamia near where the reservoir of the Kralkızı Dam now lies, around 50km north of the modern Turkish city of Diyarbakır (ancient Amida). 193 The third mention of the sources of the Tigris does not mention underground passages or Chalonitis. In book 11, Strabo locates the sources of the Tigris (τοῦ Τίγριος πηγαὶ) in the Niphates (Νιφάτης) Mountains, which he locates between the Taurus proper and the 189 The course of the two rivers is described previously on two occasions, at 11.12.3 (in his description of the Taurus) and 11.14.2 (in his desription of Armenia). In both, the Euphrates is described in much more detail: at 11.12.3, Strabo emphasises the primacy of the Euphrates in terms of its size and length and the Tigris is barely mentioned at 11.14.2. In both places he emphasises the winding course of the Euphrates and that the two rivers form Mesopotamia by enclosure: “ὁ δὲ Τίγρις ... συνάπτει τῷ Εὐφράτῃ πλησίον καὶ ποιεῖ τὴν Μεσοποταμίαν πρὸς αὐτόν” (11.12.3, “The Tigris... comes close to the Euphrates and with it forms Mesopotamia”); Ὁ γὰρ Εὐφράτης ... ποιεῖ τὴν Μεσοποταμίαν πρὸς τὸν Τίγριν” (11.14.2, “For the Euphrates... forms Mesopotamia with the Tigris”). 190 For the rivers and lake discussed in this section, see Map 2: Strabo's Assyria. 191 Chalonitis: Strabo 11.14.8. Gordyene: Strabo 16.1.21. 192 See Appendix 1.1. 193 For the geomorphology of the upper Tigris catchment area, see Nicoll (2010) 411–12. The extensive water manipulation of the Tigris and Euphrates under the auspices of the GAP (Güneydoğu Anadolu Projesi) Southeast Anatolian regional development project, has altered many aspects of the landscape of my study area, most notably the submergence of ancient sites along the rivers (Zeugma and Samosata in particular), but also in the lowering of the water level of Lake Hazar below the former level of the natural underground outflow into the Tigris. The lake now drains into the Tigris through artificial tunnels. 60/448 Zagros. 194 Strabo elsewhere places Mount Niphates “above Mount Masius” (which in turn is above Nisibis) “far towards the east, opposite Gordyene”, 195 while a third reference places Niphates to the north of Media. 196 These mountains should thus fall somewhere around a line between Lake Van and Lake Urmia. 197 It is likely that when Strabo refers to the source of the Tigris, he is referring to one of that river’s major tributaries, such as the Botan river which arises near Lake Van and runs east into the Tigris. 198 Classical authors knew the Botan as the Kentrites, a substantial river which “divided Armenia and the country of the Carduchi”. 199 Strabo’s understanding of the source of the Tigris is important for determining his conception of “the Taurus”, especially with respect to the borders of Mesopotamia. In the north, the border of Mesopotamia is Armenia and the Taurus Mountains. Exactly where in 194 Strabo 11.12.4: “τὰ μὲν δὴ πρόσβορρα ὄρη οὕτω καλοῦσι, τὰ δὲ νότια τὰ πέραν τοῦ Εὐφράτου τῆς Καππαδοκίας καὶ τῆς Κομμαγηνῆς πρὸς ἕω τείνοντα κατ’ ἀρχὰς μὲν αὐτὸ τοῦτο καλεῖται Ταῦρος, διορίζων τὴν Σωφηνὴν καὶ τὴν ἄλλην Ἀρμενίαν ἀπὸ τῆς Μεσοποταμίας· τινὲς δὲ Γ ορδυαῖα ὄρη καλοῦσιν. ἐν δὲ τούτοις ἐστὶ καὶ τὸ Μάσιον, τὸ ὑπερκείμενον τῆς Νισίβιος ὄρος καὶ τῶν Τιγρανοκέρτων. ἔπειτα ἐξαίρεται πλέον καὶ καλεῖται Νιφάτης·ἐνταῦθα δέ που καὶ τοῦ Τίγριος πηγαὶ κατὰ τὸ νότιον τῆς ὀρεινῆς πλευρόν· εἶτ’ ἀπὸ τοῦ Νιφάτου μᾶλλον ἔτι καὶ μᾶλλον ἡ ῥάχις ἐκτεινομένη τὸ Ζάγριον ὄρος ποιεῖ τὸ διορίζον τὴν Μηδίαν καὶ τὴν Βαβυλωνίαν.” (The mountains on the north, then, bear these names, whereas those on the south, on the far side of the Euphrates, in their extent towards the east from Cappadocia and Commagene, are, at their beginning called Taurus proper, which separates Sophene and the rest of Armenia from Mesopotamia; by some these are called the Gordyaean Mountains, and among these belongs also Masius, the mountain which is situated above Nisibis and Tigranocerta. Then the Taurus rises higher and bears the name Niphates; and somewhere here are the sources of the Tigris, on the southern side of the mountainous country. Then from the Niphates the mountain-chain extends still farther and farther and forms the mountain Zagrus which separates Media and Babylonia.) Horace uses the Niphates as an indicator of the Parthian lands over which Augustus celebrated a triumph in BCE (Carmina 2.9.20). 195 Strabo 11.14.2: “τοῦ δὲ Μασίου ὑπέρκειται πρὸς ἕω πολὺ κατὰ τὴν Γορδυηνὴν ὁ Νιφάτης”. 196 Strabo 11.13.3. 197 Both Pliny (NH 5.98) and Pomponius Mela (1.81) give imprecise locations for Niphates amidst lists of mountain ranges comprising the Taurus. Niphates is identified with the modern Tendürük Dağı northeast of Lake Van, Radt (2002) 7.306. 198 Tozer identified the “Lake Thopitis” described by Strabo in 11.14.8 as Lake Van, (1897) 268; (1881) 293; Dillemann (1962) 41–42. 199 Xen. Anab. 4.3.1: Ταύτην δ’ αὖ τὴν ἡμέραν ηὐλίσθησαν ἐν ταῖς κώμαις ταῖς ὑπὲρ τοῦ πεδίου παρὰ τὸν Κεντρίτην ποταμόν, εὖρος ὡς δίπλεθρον, ὃς ὁρίζει τὴν Ἀρμενίαν καὶ τὴν τῶν Καρδούχων χώραν. καὶ οἱ Ἕλληνες ἐνταῦθα ἀνέπνευσαν ἄσμενοι ἰδόντες πεδίον·ἀπεῖχε δὲ τῶν ὀρέων ὁ ποταμὸς ἓξ ἢ ἑπτὰ στάδια τῶν Καρδούχων. (For that day again they found quarters in the villages that lie above the plain bordering the Kentrites river, which is about two plethra in width and separates Armenia from the country of the Carduchians. There the Greeks took breath, glad to behold a plain; for the river was distant six or seven stadia from the mountains of the Carduchians). Loeb trans. The same story is related by Diodorus Siculus 14.27.7. 61/448 this range Strabo considered the border to be is unclear. 200 He does not explicitly define a border as does Xenophon between Armenia and the Carduchians at the Kentrites. 201 To complicate matters, Strabo repeatedly locates the important cities of Nisibis and Tigranocerta relative to Mount Masius. 202 Because determining the location of Masius requires a detailed examination of the locations of those two cities, both of which Strabo places in the sub-division of Mesopotamia called Mygdonia, as well as a discussion of the location of another Mesopotamian sub-division, Gordyene, we shall first examine how Strabo divides space within Mesopotamia, then return to the question of the region’s northern boundary . This definition of Mesopotamia according to geophysical features of the space contrasts with the ethnic basis of his initial overview of οἱ Ἀσσύριοι at the start of book 16. As we shall see below , Strabo’s detailed geographical narrative mixes both methods of defining and describing Mesopotamia and the spaces and places within it. Following his broad initial description of the regional topology , Strabo divides Mesopotamia into three sections, describing each in turn. At the beginning of book 16, Strabo promised to describe tribes of Mesopotamia near the Gordyaeans (τὰ τῆς Μεσοποταμίας ἔθνη τὰ περὶ Γ ορδυαίους), the Mygdonians around Nisibis as far as Zeugma (τοὺς περὶ Νίσιβιν Μυγδόνας μέχρι τοῦ Ζεύγματος τοῦ κατὰ τὸν Εὐφράτην), and the Arabians on the other side of the Euphrates (τῆς πέραν τοῦ Εὐφράτου πολλὴ ἣν Ἄραβες κατέχουσι). 203 Although he changes the order, Strabo’s narrative does indeed divide Mesopotamian between these three spaces. Mygdonia Having defined the space at a broad level, Strabo proceeds through a series of closer examinations of 200 On the difficulty of assigning precise borders in mountainous areas, see Sahlins (1989). 201 Xen. Anab. 4.3.1. 202 Strabo 11.12.4: “τινὲς δὲ Γ ορδυαῖα ὄρη καλοῦσιν. ἐν δὲ τούτοις ἐστὶ καὶ τὸ Μάσιον, τὸ ὑπερκείμενον τῆς Νισίβιος ὄρος καὶ τῶν Τιγρανοκέρτων.” (Some call these the Gordyaean mountains among which is Masius, the mountain which lies above Nisibis and Tigranocerta.) The location and identity of Mount Masius is discussed below . 203 Strabo 16.1.1 (quoted above, p.41). 62/448 the subdivided area (See Map 3: Strabo’s Syria). He begins with ἡ παρόρειος, the part near the mountains, which includes at least part of the land alongside the Euphrates. Ἔστι δ’ ἡ μὲν παρόρειος εὐδαίμων ἱκανῶς· ἔχουσι δ’ αὐτῆς τὰ μὲν πρὸς τῷ Εὐφράτῃ καὶ τῷ ζεύγματι, τῷ τε νῦν τῷ κατὰ τὴν Κομμαγηνὴν καὶ τῷ πάλαι τῷ κατὰ τὴν Θάψακον, οἱ Μυγδόνες κατονομασθέντες ὑπὸ τῶν Μακεδόνων· ἐν οἷς ἐστιν ἡ Νίσιβις, ἣν καὶ αὐτὴν Ἀντιόχειαν τὴν ἐν τῇ Μυγδονίᾳ προσηγόρευσαν, ὑπὸ τῷ Μασίῳ ὄρει κειμένην, καὶ Τιγρανόκερτα καὶ τὰ περὶ Κάρρας καὶ Νικηφόριον χωρία καὶ Χορδίραζα καὶ Σίννακα, ἐν ᾗ Κράσσος διεφθάρη, δόλῳ ληφθεὶς ὑπὸ Σουρήνα τοῦ τῶν Παρθυαίων στρατηγοῦ. 204 The name of the people, Mygdones (οἱ Μυγδόνες), and thus the area, Mygdonia, derived from the early Hellenistic colonisation of the region by the Macedonians. Strabo reports that the Mygdones were named by the Macedonians (κατονομασθέντες ὑπὸ τῶν Μακεδόνων), but it is unclear how they came to apply this particular name to these people. Three modes of naming are attested within the Greek and Macedonian colonial project: the imposition of an entirely new name, the transliteration of an existing name, or the translation of a non-Greek word into Greek. This latter mode could include the identification of an eponymous founder from Greek mythology with some aspect of local culture or topography , as may be the case here. 205 The Greek name Mygdones was probably derived in some way from the Phrygian hero named Mygdon who appears in the Iliad. 206 While it is implied, it is not precisely clear whether Strabo’s ἡ παρόρειος and the land he assigns to οἱ Μυγδόνες are the same. The descriptions of the two spaces are not explicitly linked by the grammar of the passage (the list of cities is linked to Mygdonia by ἐν οἷς ἐστιν), but Strabo supplies no additional subdivisions of ἡ παρόρειος. Moreover, his scant depiction of ἡ παρόρειος – it is εὐδαίμων ἱκανῶς – 204 Strabo 16.1.23: “The land near the mountains is sufficiently fertile. The Mygdonians (as the Macedonians call them) hold the lands near the Euphrates and Zeugma, both the current Zeugma in Commagene and the old Zeugma near Thapsacus. In Mygdonia are Nisibis, also called Mygdonian Antioch, which lies under Mount Masius, as well as Tigranocerta, the lands of Carrhae and Nikephorion, Chordiraza and Sinnaca, where Crassus was killed, seized treacherously by Suren the Parthian general.” 205 Pigulevskaja (1963) 51, suggests that “Mygdonia” is derived from the Syriac magda (ܐܕܓܡ, “fruit”). 206 Hom. Il. 3.186. There are many other examples to this throughout Asia. In close proximity , note the implicit link between the hero Gordys and the region Gordyene at Strabo 16.1.25: “Λέγεται δὲ Γόρδυς ὁ Τριπτολέμου τὴν Γ ορδυηνὴν οἰκῆσαι...” (Gordys, the son of Triptolemus, is said to have taken up his abode in Gordyenê...) 63/448 matches the characterisation of Mygdonia in other sources. Describing a gift of part of Armenia to a Parthian king, Josephus notes that the country around Nisibis is ἀγαθός, surely a reference to its fertility . 207 In Polybius' account of Antiochus III's march across Mesopotamia against the rebellious satrap Molon, 207 Josephus Ant. Iud. 20.68: “ἔδωκεν δὲ καὶ χώραν πολλὴν αὐτῷ κἀγαθὴν τοῦ τῶν Ἀρμενίων βασιλέως. Νίσιβις δέ ἐστιν ὄνομα τῇ γῇ, καὶ ἐν αὐτῇ πρότερον Μακεδόνες ἐκτίσαντο πόλιν Ἀντιόχειαν, ἣν Ἐπιμυγδονίαν προσηγόρευσαν.” (He furthermore gave him an extensive and productive territory which he carved from that of the king of Armenia. The district is called Nisibis, and in it the Macedonians had in days of old founded the city of Antioch which they surnamed Epimygdonia.) Loeb trans. Josephus is identifies the city as Antioch Epimygdonia (i.e. Mygdonian Antioch) and the district around it as Nisibis. This confusion of city and surrounding area is frequently seen for Batnae and Anthemusia, see Appendix 1.3. Cassius Dio (68.26.1) mentions the forests around Nisibis which existed in Trajan’s time. Ammianus Marcellinus (19.9.5) also mentions forests between Samosata and Nisibis in the fourth century. 64/448 Map 5: Strabo's Mygdonia Antiochus and his army sit out the worst of the winter for 40 days at Nisibis. This implies the ability to store substantial quantities of food probably drawn from a fertile hinterland. 208 Finally , in his campaign against Tigranes, Lucullus had defeated the Armenian forces blocking the road to their capital at Artaxata (near modern Artashat), but bad weather and discontent in the ranks forced him to turn south to Mesopotamia. The army descended into Mygdonia, a χώραν παμφόρον καὶ ἀλεεινὴν (“an area productive and open to the sun”). 209 The cities and places included by Strabo in Mygdonia are not related in a geographical order; however, it is not a coincidence that the first cities that Strabo lists in his description of the region are the two Zeugmas. Strabo describes the Mygdonians as possessing the parts near the Euphrates and the two Zeugmas (τῷ νῦν Zeugma near Commagene, and τῷ πάλαι Zeugma near Thapsacus), both of which lie on the Euphrates. 210 The town usually referred to as Zeugma today lay west of the Euphrates. It was also known as Seleucia on the Euphrates, and was paired with a twin on the east bank, Apamaea. In his description of Commagene, Strabo refers to Seleucia as a “fortress of Mesopotamia”. 211 Strabo mentions Zeugma/Seleucia in the context of geographical areas on both sides of the Euphrates (Commagene and Mygdonia). For Strabo, the precise location of river crossings on either river bank seems irrelevant to his conception of their location. He understands them as crossing points over the river which share and connect 208 Polybius 5.51.1. See Chapter 5. 209 Plut. Luc. 32.4: “ὡς δ’ οὐκ ἔπειθεν, ἀπῆγεν αὐτοὺς ὀπίσω, καὶ κατ’ ἄλλας ὑπερβολὰς διελθὼν τὸν Ταῦρον, εἰς τὴν λεγομένην Μυγδονικὴν κατέβαινε, χώραν παμφόρον καὶ ἀλεεινὴν καὶ πόλιν ἐν αὑτῇ μεγάλην καὶ πολυάνθρωπον ἔχουσαν, ἣν οἱ μὲν βάρβαροι Νίσιβιν, οἱ δ’ Ἕλληνες Ἀντιόχειαν Μυγδονικὴν προσηγόρευον.” (But since he could not persuade them, he led them back, and crossing the Taurus by another pass, descended into the country called Mygdonia, which is fertile and open to the sun, and contains a large and populous city , called Nisibis by the Barbarians, Antioch in Mygdonia by the Greeks.) Loeb trans. 210 For the locations mentioned in the subsequent paragraphs, see Map 3: Strabo’s Syria. The location of Thapsacus is discussed in Appendix 1.18; the many potential locations for the old Zeugma at Thapsacus are marked on Map 2: Strabo's Assyria. Nor does the location of both Zeugmata in Mygdonia help narrow a location for Thapsacus; Mygdonia includes cities lying at least as far down the Euphrates as Nikephorion at any rate, so even placing Thapsacus near Zenobia would scarcely be more of a stretch. For the location of Sinnaca, see Appendix 1.17. Ptolemy lists some of the cities mentioned in Strabo, but he does not identify Mygdonia among the chora of Mesopotamia. 211 Strabo 16.2.3, see above, p.45. 65/448 both banks. This conception is based more in the praxis and experience of a river crossing than a precise topographical location. It is also entirely fitting with Strabo’s name for the site, Zeugma, crossing. The crossing points over the Euphrates were critical points of contact with important trade and military routes between Roman Syria and Parthian Babylonia, as their use by Alexander and Cyrus the Younger show . 212 The remaining places that Strabo assigns to the Mygdonians are Nisibis, Tigranocerta, Carrhae, Nikephorion, Chordiraza and Sinnaca. Three of these places are cities with well known locations: Nisibis is modern Nusaybin; Carrhae is ancient Harran, near modern Altinbasak, 44 km southeast of Şanlıurfa (ancient Edessa); and the remains of Nikephorion are near the modern Syrian town of Ar-Raqqa. 213 Of those three cities, only Nikephorion is on the Euphrates, but almost 200km from the Taurus mountains, while Nisibis is a similar distance from the Euphrates, but quite close to the mountains – very close if the Tur Abdin is included. 214 Two of the other locations, Chordiraza and Sinnaca both lie in northeast Mesopotamia: the former near the Euphrates and the latter near Carrhae. 215 It may seem that these distances stretch the definition of παρόρειος and πρὸς τῷ Εὐφράτῃ, but since Strabo is explicit that Nisibis belongs to the Mygdonians (he calls it Mygdonian Antiocheia), despite the distance from the river, it is permissible to grant the same loose application of proximity to the mountains to the rest of the Mygdonian lands. Moreover, if ἡ παρόρειος and Mygdonia are the same, then Strabo has already defined the Mygdonians as inhabiting lands near the mountains as well as πρὸς τῷ Εὐφράτῃ. The location of the final site mentioned by Strabo is much disputed by modern scholars. The remains of Tigranocerta, the first-century BCE capital of Tigranes’ short-lived Armenian empire, have not been 212 On Thapsacus, see Appendix 1.18. Routes of commercial movement across Mesopotamia will be fully discussed in Chapter 5. 213 Nisibis: Radt (2002) 8.278; Kessler, Karlheinz, ‘Nisibis’, BNP 9.777-79. Carrhae: Radt (2002) 8.279; Kessler, Karlheinz, ‘Harran’ BNP 5.1152. Nicephorium: Radt (2002) 8.279; Kessler, Karlheinz, ‘Nicephorium’ BNP 9.714. For more, see Appendices 1.13, 1.6 and 1.1.13 respectively. 214 Lehmann-Haupt (1910) 1.370 suggests an etymological relationship between Tur and Taurus. Radt (2002) 7.306. 215 For more on both sites, see Appendices 1.7 and 1.17. 66/448 securely identified despite the relative abundance of literary records mentioning it. 216 Three sites in particular have been proposed: Tell Ermen (modern Kızıltepe) south of the Tur Abdin and west of Nisibis; late-antique Martyropolis (modern Meiafarkin, now usually known by its Turkish name, Silvan) north of the Tigris at the foot of the Taurus; and at a set of badly decayed ruins at Arzan, also north of the Tigris, but further east on the banks of the Garzan-su. Those sources which specifically mention the location of Tigranocerta relative to the Tigris place the city to the north, between the Tigris and the Taurus range proper, while several other sources place the city in Armenia, which implies a site north of the Tigris. 217 This casts serious doubt on Tell Ermen on the Mesopotamian plain near Nisibis. Literary descriptions of the site and the nearby river in Tacitus, Pliny and Plutarch render Martyropolis an unlikely candidate. 218 This leaves Arzan, which matches the various descriptions best, but not perfectly; there are some discrepancies with Plutarch’s account. Nevertheless, it remains the most likely candidate. This likely location of Tigranocerta on the north site of the Tigris has aroused consternation in scholars attempting to make sense of Strabo’s placement of the city in Mesopotamia. This difficulty arises because Strabo says that Tigranocerta lay in Mesopotamia and that Mesopotamia ended at the Tigris. However, as we have seen, Strabo seems to have had a different conception of what constituted the Upper Tigris. Furthermore, the intersection between the Tigris and the Taurus confounds other attempts to rigidly apply Strabo’s delimitations. In his description of Mesopotamia, Strabo describes Nisibis as located “under” Mount Masius (ὑπὸ τῷ Μασίῳ ὄρει κειμένην). Mount Masius is usually identified with the low foothills north of Nisibis which separates the plains of Mesopotamia from the upper Tigris river valley called the Tur Abdin, and the Karakaš Dağı to the west of those hills. 219 In his description of Armenia, Strabo places both 216 A full discussion of the literary records, the proposed sites, and the arguments for each can be found in Appendix 1.19. 217 Relative to the Tigris: Ptol. Geog. 5.13.22; Tab. Peut. 10B3; Orosius 6.3.6; Plut. Luc. 24.7. In Armenia: Pliny NH. 6.26; Eutropius 6.9; Buzandaran Patmut'iwnk' 4.24. 218 Tac. Ann. 15.4; Pliny NH 6.129; Plut. Luc. 27. 219 Radt (2002) 7.306; Dillemann (1962) 32–34; Syme (1995) 47. Syme calls them “a series of undulations rather 67/448 Nisibis and Tigranocerta below Mount Masius, which he in turn places among the mountains of Gordyene. 220 Syme objects to this characterisation of Masius as part of the Gordyaean mountains: “the Gordyaean massif is... the easterly continuation of the Taurus... divided from Mesopotamia by the River Tigris... Yet Strabo here puts the Gordyaean mountains west of the Tigris.” 221 For Syme, the Tigris is a geological and onomastic barrier: the Gordyaean mountains must end at the Tigris. Any continuation of the mountains on the other side of the river must have another name and be a geographically distinct entity . This is unnecessarily rigid. The course of the upper Tigris is a ravine not unlike the course of the upper Euphrates between Commagene and Cappadocia. Ancient writers had no problem describing that latter river forcing its way through the Taurus without forcing a conceptual division between the Taurus to the east of the Euphrates and the Taurus to the west. Gordyene Next in Strabo’s presentation comes the second of the three parts into which he divides Mesopotamia, “the places of the Gordyaeans, near the Tigris”: Πρὸς δὲ τῷ Τίγρει τὰ τῶν Γ ορδυαίων χωρία οὓς οἱ πάλαι Καρδούχους ἔλεγον, καὶ αἱ πόλεις αὐτῶν Σάρεισά τε καὶ Σάταλκα καὶ Πίνακα, κράτιστον ἔρυμα, τρεῖς ἄκρας ἔχουσα, ἑκάστην ἰδίῳ τείχει τετειχισμένην, ὥστε οἷον τρίπολιν εἶναι. ἀλλ’ ὅμως καὶ ὁ Ἀρμένιος εἶχεν ὑπήκοον καὶ οἱ Ῥωμαῖοι βίᾳ παρέλαβον, καίπερ ἔδοξαν οἱ Γ ορδυαῖοι διαφερόντως ἀρχιτεκτονικοί τινες εἶναι καὶ πολιορκητικῶν ὀργάνων ἔμπειροι· διόπερ αὐτοῖς εἰς ταῦτα ὁ Τιγράνης ἐχρῆτο. ἐγένετο δὲ καὶ ἡ λοιπὴ Μεσοποταμία ὑπὸ Ῥωμαίοις, Πομπήιος δ’ αὐτῆς τὰ πολλὰ τῷ Τιγράνῃ προσένειμεν ὅσα ἦν ἀξιόλογα· ἔστι γὰρ εὔβοτος ἡ χώρα καὶ εὐερνὴς ὥστε καὶ τὰ ἀειθαλῆ τρέφειν καὶ ἄρωμα τὸ ἄμωμον· καὶ λεοντοβότος ἐστί· φέρει δὲ καὶ τὸν νάφθαν καὶ τὴν γαγγῆτιν λίθον, ἣν φεύγει τὰ ἑρπετά. 222 than a mountain chain”. W eissbach “Masion” RE Halb. 28, cols.2068-69. 220 Strabo 11.12.4: “τινὲς δὲ Γ ορδυαῖα ὄρη καλοῦσιν. ἐν δὲ τούτοις ἐστὶ καὶ τὸ Μάσιον, τὸ ὑπερκείμενον τῆς Νισίβιος ὄρος καὶ τῶν Τιγρανοκέρτων.” (Some call these the Gordyaean mountains among which is Masius, the mountain which lies above Nisibis and Tigranocerta.) 221 Syme (1995) 48. 222 Strabo 16.1.24: “Near the Tigris are the places of the Gordyaeans, who were once called Carduchians. Their cities are Sareisa, Satalka and Pinaka, a very strong fortress with three citadels, each fortified by its own walls so that it is a kind of triple city. But the Armenian [king] held it as a subject and the Romans seized it by force, although the Gordyeni had a particular reputation as builders and were skilled in the technologies of siege warfare; for this reason Tigranes employed them in this capacity. The rest of Mesopotamia also came under 68/448 There are three parts to Strabo’s description of Gordyene: a geographical description of the location and contents of the space, a brief overview of political control, and a description of the flora and fauna of the area. Strabo places the territory of the Gordyaeans (τὰ τῶν Γ ορδυαίων χωρία) next to the Tigris (πρὸς δὲ τῷ Τίγρει) and within Mesopotamia. This implies that these people live on the right bank of the river, as in fact, Strabo makes explicit in his book on Armenia. ἐκεῖθεν δ᾽ ἤδη πρὸς τὴν Ὦπιν καὶ τὸ τῆς Σεμιράμιδος καλούμενον διατείχισμα ἐκεῖνός τε καταφέρεται τοὺς Γ ορδυαίους ἐν δεξιᾷ ἀφεὶς καὶ τὴν Μεσοποταμίαν ὅλην, καὶ ὁ Εὐφράτης τοὐναντίον ἐν ἀριστερᾷ ἔχων τὴν αὐτὴν χώραν: πλησιάσαντες δὲ ἀλλήλοις καὶ ποιήσαντες τὴν Μεσοποταμίαν... 223 Strabo is perfectly clear: the Gordyaeans and Mesopotamia are on the right (west) bank of the Tigris and lie in the land between that river and the Euphrates. 224 The only problem is that the country of the Gordyaeans is well-known to include parts of the Taurus mountains to the east (i.e. on the left bank) of the Tigris. Strabo is explicit that the Gordyaeans are the same people as the Carduchians (οὓς οἱ πάλαι Καρδούχους ἔλεγον, “whom those in earlier times were called the Carduchians”). 225 The Carduchians are most famous now , as they probably were also in Strabo’s time, for their appearance in Xenophon’s Anabasis as warlike mountaineers, where they occupy a stretch of the Taurus east of the Tigris and north of the Assyrian heartland whose ruins Xenophon describes. 226 Elsewhere in Strabo’s Geography , these mountains which Roman control and Pompey assigned to Tigranes most of this region—those parts worth mentioning. For the area is rich in pasturage and flourishing, bearing both evergreens and the spice amomum [an aromatic spice related to cardamon]; it is also a feeding ground for lions; and it produces naptha and the stone gangitis, which reptiles avoid.” 223 Strabo 11.14.8: “From there [the Tigris] flows down towards Opis and the so-called wall of Semiramis, passing by the Gordyaeans on the right and all of Mesopotamia. The Euphrates, on the contrary , has the same country on the left. Approaching each other, they form Mesopotamia.” 224 See also the description of Tigranes’ conquests at Strabo 11.14.15: “αὐξηθεὶς... ὑπηκόους δ᾽ ἔσχε καὶ τὸν Ἀτροπατηνὸν καὶ τὸν Γ ορδυαῖον, μεθ᾽ ὧν καὶ τὴν λοιπὴν Μεσοποταμίαν, ἔτι δὲ τὴν Συρίαν αὐτὴν καὶ Φοινίκην διαβὰς τὸν Εὐφράτην ἀνὰ κράτος εἷλεν.” (But when he had grown in power... he subjugated to himself the rulers of Atropene and Gordyaea, and along with these the rest of Mesopotamia, and also crossed the Euphrates and by main strength took Syria itself and Phoenicia.) 225 Pliny NH 6.44; Radt (2002) 8.280. 226 Xen. Anab. 3.5-4.3; Diodorus 14.27.4; Syme (1995) 53. On Assyrian ruins: Xen. Anab. 3.4. 69/448 extend on both sides of the Tigris river share their name with the Gordyaean people. 227 Pliny also places the Gordyaeans east of the Tigris. 228 Since Xenophon, Pliny and Ptolemy all place the Gordyaeans/Carduchians on the east bank of the Tigris, it seems highly likely that Strabo conceived of them as inhabiting the same space, and thus dwelling on both banks of the Tigris. Just as we saw previously in the example of the name Masius which could be applied to a continuous range on both sides of the Tigris, so in Strabo’s understanding of the region, Gordyene extends on both sides of the Tigris, and lies within Mesopotamia. 227 Strabo considers Masius/Tur Abdin to be part of the Gordyaean mountains, above, p.45. 228 Pliny (NH 6.44) reports that the Adiabeni were shared a common border with the Gordyeni: “Adiabenis conectuntur Carduchi quondam dicti, nunc Cordueni, praefluente Tigri...” (The Carduchi, as they were once called, are connected to the Adiabeni. Now they are called the Cordueni [Gordyeni]. The Tigris flows past them.) 70/448 Map 6: Gordyene The cities and fortresses named by Strabo are too insecurely located to contribute to our understanding of Strabo’s conception of Gordyene. The long-standing equation of Pinaka (Πίνακα) with the modern village of Finek has been confirmed in recent decades, 229 but the locations of Sareisa (Σάρεισά) and Satalka (Σάταλκα) are unknown. Three possible identifications for Sareisa have been suggested (Map 6: Gordyene): The Barrington Atlas suggests Şırnak in the Taurus/Gordyaean Mountains, northeast of Cizre; 230 Syme suggests that the site be located near the confluence of the Bohtan and the Tigris; 231 Dillemann calls the identification of Sareisa with “Charich” (Turkish Gercüş) in the Tur Abdin north of Nisibis plausible and offers Chattakh as an identification of Satalka. 232 None of these suggestions are supported with significant argument. The location and identification of Sareisa must await archaeological exploration of these sites. To return briefly to the question of the northern boundary of Strabo’s Mesopotamia. Strabo’s inclusion of Tigranocerta and Gordyene within “Mesopotamia” shows that in his understanding, Mesopotamia extended past the Tur Abdin to include the Upper Tigris Basin and even extended a short distance into the Taurus mountains. When he notes that the northern boundary is the Taurus, Strabo is careful to mention that the adjacent region is Armenia. 233 Strabo most likely considered the northern-facing slopes and watersheds of the Taurus part of Armenia and the southern-facing slopes and watersheds part of 229 For example: Dillemann (1962) 84; Syme (1995) 30 n.9, 54. Strabo describes Pinaka as a strongly fortified site with three citadels (“κράτιστον ἔρυμα, τρεῖς ἄκρας ἔχουσα, ἑκάστην ἰδίῳ τείχει τετειχισμένην, ὥστε οἷον τρίπολιν εἶναι”). Ammianus (20.7.1) reports that the late antique fortress of Bezabde was originally called Phaenicha: “...Bezabden, quam Phaenicham quoque institutores veteres appellarunt...” (...Bezabde, which its ancient founders once called Phaenicha...). While it initially seems unproblematic that Pinaka, Phaenicha and Finek should be the same place, the site seemed not to match Strabo’s description, Dillemann (1962) 84. However, based on a surface survey of the area, including hills on both sides of the Tigris, Algaze has identified the twin villages of Eskiyapı (Fenik) and Hendekköy which lie on the Tigris about 13 km north of Cizre with Strabo’s Pinaka and the Phaenicha/Bezabde of Ammianus Marcellinus, Algaze (1989) 247–252. 230 BAtlas 89 E3; Kroll, S., M. Roaf, St J. Simpson, T . Sinclair, R. Talbert, T . Elliott, S. Gillies. "Places: 874682 (Sareisa)". Pleiades. <http://pleiades.stoa.org/places/874682> [Accessed: November 13, 2012 12:21 pm] 231 Sarisa may be the same place as the Še-ri-eš-še conquered by Tiglath-Pileser I, Syme (1995) 54 n.32. 232 Dillemann (1962) 111; Matthews (1989) 50–51. Dillemann locates Chattakh deeper in the Taurus Mountains, on the bank of the Botan south of Lake Van. 233 Strabo 16.1.21: “πρὸς ἄρκτον δὲ ὁ Ταῦρος ὁ τοὺς Ἀρμενίους διορίζων ἀπὸ τῆς Μεσοποταμίας.” (Quoted and translated above, p.44). Strabo also refers to the Taurus as the boundary between Armenia and Mesopotamia in his description of Armenia (11.12.2). 71/448 Mesopotamia. Southern Parts The third and final sub-division of Strabo’s Mesopotamia comprises the southern parts of Mesopotamia, τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα τῆς Μεσοποταμίας καὶ ἀπωτέρω τῶν ὀρῶν, “the parts of Mesopotamia towards the south and further from the mountains”. In this section, Strabo refers to several cities in the area and settled communities along the Euphrates (see Map 5: Strabo's Mygdonia). His text raises several problems questions about the relationship between these settled communities and the spaces he describes. I will discuss these problems after considering the most prominent population of the area, the Skenitai. τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα τῆς Μεσοποταμίας καὶ ἀπωτέρω τῶν ὀρῶν ἄνυδρα καὶ λυπρὰ ὄντα ἔχουσιν οἱ σκηνῖται Ἄραβες, λῃστρικοί τινες καὶ ποιμενικοί, μεθιστάμενοι ῥᾳδίως εἰς ἄλλους τόπους, ὅταν ἐπιλείπωσιν αἱ νομαὶ καὶ αἱ λεηλασίαι. τοῖς οὖν παρορείοις ὑπό τε τούτων κακοῦσθαι συμβαίνει καὶ ὑπὸ τῶν Ἀρμενίων· ὑπέρκεινται δὲ καὶ καταδυναστεύουσι διὰ τὴν ἰσχύν· τέλος δ’ ὑπ’ ἐκείνοις εἰσὶ τὸ πλέον ἢ τοῖς Παρθυαίοις· ἐν πλευραῖς γάρ εἰσι κἀκεῖνοι τήν τε Μηδίαν ἔχοντες καὶ τὴν Βαβυλωνίαν. 234 The pastoralist people that Strabo called Skenitai (tent-dwelling) Arabs dwell in these southern parts of Mesopotamia away from the mountains (ἀπωτέρω τῶν ὀρῶν). 235 Although Strabo characterises their land as barren and waterless (ἄνυδρα καὶ λυπρὰ ὄντα), much of it is not desert. Rather, it is marginal land capable of supporting pastoralism: the Skenitai are ποιμενικοί (shepherds) who move about between pasture- lands (αἱ νομαὶ). It is not possible to fix a line precisely between the regions to the north which support dryland farming and the regions to the south capable only of nomadic pastoralism, but it probably lies 234 Strabo 16.1.26: “The Skenitai Arabs occupy those parts of Mesopotamia which incline towards the south and are farther from the mountains, and are waterless and barren. [The Skenitai] are bandits and shepherds, who readily move from place to place when pasture and booty fail them. As a result, the people near the mountains are harassed by the Skenitai and also by the Armenians, who dwell above them and oppress them through brute force. They are subject mostly to the Armenians or else to the Parthians, who, possessing both Media and Babylonia, are situated on the sides of the country.” 235 The Skenitai will be discussed in detail in Chapter 5. 72/448 somewhere around the 200mm isohyet (Map 7: The 200mm Isohyet). 236 This line represents an average rainfall of 200mm per annum. Above this line, dryland farming (i.e. farming dependent on irrigation) is reliable, below this line, it is not. 237 Pastoralists follow traditional routes between pasture-lands as periodic rainfall and grass growth permits. Settled people did dwell in these arid regions, especially along the rivers. Strabo mentions such people in his description of the main route through Mesopotamia between the Mediterranean and Babylonia, as will be discussed in chapter five. Navigation of this route was mediated by the Skenitai, but there is evidence of several settled communities along the way . Strabo writes: Μεταξὺ δὲ τοῦ Εὐφράτου καὶ τοῦ Τίγριος ῥεῖ καὶ ἄλλος ποταμὸς Βασίλειος καλούμενος, καὶ 236 Map from Kennedy and Riley (1990) 25, Figure 2: Physical Geography of the Middle East. Note that this map does not include the Balikh which enters the Euphrates at about the point where the 200mm isohyet crosses the latter river. Cf. Map 8: The Skenitai and their Lands. 237 The difference in plant growth and the appearance of the landscape varies significantly according to the seasons. 73/448 Map 7: The 200mm Isohyet περὶ τὴν Ἀνθεμουσίαν ἄλλος Ἀβόρρας· διὰ δὲ τῶν Σκηνιτῶν, ὑπὸ τῶν Μαλίων νυνὶ λεγομένων, καὶ τῆς ἐκείνων ἐρημίας ἡ ὁδὸς τοῖς ἐκ τῆς Συρίας εἰς Σελεύκειαν καὶ Βαβυλῶνα ἐμπορευομένοις ἐστίν. ἡ μὲν οὖν διάβασις τοῦ Εὐφράτου κατὰ τὴν Ἀνθεμουσίαν ἐστὶν αὐτοῖς, τόπον τῆς Μεσοποταμίας· ὑπέρκειται δὲ τοῦ ποταμοῦ σχοίνους τέτταρας διέχουσα ἡ Βαμβύκη, ἣν καὶ Ἔδεσσαν καὶ Ἱερὰν πόλιν καλοῦσιν, ἐν ᾗ τιμῶσι τὴν Συρίαν θεὸν τὴν Ἀταργάτιν. διαβάντων δὲ ἡ ὁδός ἐστι διὰ τῆς ἐρήμου ἐπὶ τοὺς τῆς Βαβυλωνίας ὅρους μέχρι Σκηνῶν, ἀξιολόγου πόλεως ἐπί τινος διώρυγος ἱδρυμένης... διέχουσι δὲ τῆς Σελευκείας αἱ Σκηναὶ σχοίνους ὀκτωκαίδεκα. 238 The rivers mentioned here by Strabo are the Balikh (or Balissos) and the Khabur. Strabo calls the former river “Basileios” (ἄλλος ποταμὸς Βασίλειος καλούμενος). 239 The Balikh itself emerges from springs south of Carrhae, but it is fed by numerous sources in the hilly , western part of northern Mesopotamia around Edessa and Carrhae. It flows south to meet the Euphrates at Raqqa/Nikephorion. The second river, which Strabo calls Ἀβόρρας, is the Khabur, usually known in Greek as Χαβώρας. 240 This flows south from its catchment area in the Tur Abdin to join the Euphrates at Circesium. It is not clear whether Strabo considered the settled communities of this region as part of Mygdonia or associated them with the Skenitai. This is particularly relevant to the unnamed communities along the river, but raises the question for the named cities too. Earlier he had defined Mygdonia as including parts of Mesopotamia between the Zeugma in Commagene and the Zeugma at Thapsacus. 241 Knowing a precise location for Thapsacus might help in establishing the bounds of Strabo’s Mygdonia, but the location of other sites already provides us with much information in this regard. 242 In particular, Strabo notes that 238 Strabo 16.1.27: “Between the Euphrates and the Tigris there flows another river, called Basileius; and in the neighbourhood of Anthemusia still another, called Aborras. The road for people travelling from Syria to Seleuceia and Babylon runs through the country of the Skenitai, now called Malians by some writers, and through their desert. Such travellers cross the Euphrates near Anthemusia, a place in Mesopotamia; and above the river, at a distance of four schoinoi, lies Bambycê, which is also called Edessa and Hierapolis, where the Syrian goddess Atargatis is worshipped; for after they cross the river, the road runs through the desert to Skenai, a noteworthy city situated on a canal towards the borders of Babylonia.... Skenai is eighteen schoinoi distant from Seleuceia.” Radt (2002) 8.282. 239 By metathesis of Balissos for Basileos. Strabo is perhaps confused by the canal in Babylonia called Naarmalcha by Ammianus (amnis regum, 24.6.1; Flumen regium, 23.6.25) and Royal Canal (τὴν Βασιλικὴν διώρυχα καλουμένην) by Polybius (5.51.6), see below p.43; Radt (2002) 8.281. 240 Radt (2002) 8.281; Kühne, “Habur” BNP . Ammianus Marcellinus also knows it by this name (Abora, 23.5.1). Ptol. Geog. 5.18.3. 241 Strabo 16.1.23, see above, p.62. 242 For a discussion of the location of Thapsacus, see Appendix 1.18, with map. 74/448 Nikephorion at the confluence of the Balikh and Euphrates is part of Mygdonia. 243 Nikephorion is downstream of the probable locations for Thapsacus, although the location near Zenobia obtained by the measurements Strabo attributes to Eratosthenes would lie further to the south. The three settled communities that Strabo mentions in this passage are Anthemusia, Bambyke and Skenai. The mention of each settlement presents a different problem. Strabo refers to Anthemusia twice. 244 He describes the Khabur river as “near Anthemusia” (καὶ περὶ τὴν Ἀνθεμουσίαν ἄλλος Ἀβόρρας). A few lines latter, he refers to Anthemusia again, this time as near the 243 Strabo 16.1.23. 244 For more on Anthemusia, see Appendix 1.3. 75/448 Map 8: The Skenitai and their Lands crossing of the Euphrates (ἡ μὲν οὖν διάβασις τοῦ Εὐφράτου κατὰ τὴν Ἀνθεμουσίαν ἐστὶν αὐτοῖς). On this second occasion, he specifically describes Anthemusia as a place in Mesopotamia (τόπον τῆς Μεσοποταμίας). Strabo is correct in locating Anthemusia in Mesopotamia near a crossing of the Euphrates, but it is not near the Khabur (Ἀβόρρας). 245 It is difficult to reconcile Strabo’s two references without assuming he takes a very broad scale view . The next community Strabo mentions is Bambyke (Βαμβύκη). Strabo had previously described this city as part of his description of Cyrrhestice in Syria; its remains lie west of the Euphrates near modern Manbij (derived from the Aramaic name, Mabbog), close to at least one crossing of the Euphrates (See Map 4: The Geographical Boundaries of Strabo's Mesopotamia). 246 Aside from including the site in his description of Mesopotamia, Strabo provides information about Bambyke which otherwise conforms to the site at Manbij. He reports that Bambyke is four schoinoi “above” the river, which corresponds to the distance between Manbij and a crossing at the confluence of the Sadjur and Euphrates rivers, the nearest point on the Euphrates, 20km to the north. 247 However, Strabo also calls Bambyke by the alternate names Edessa and Hierapolis (ἣν καὶ Ἔδεσσαν καὶ Ἱερὰν πόλιν καλοῦσιν), and expands on the later by specifying the that name “holy city” refers to the cult of Atargatis (ἐν ᾗ τιμῶσι τὴν Συρίαν θεὸν τὴν Ἀταργάτιν). 248 While Hierapolis appears often in the sources, Strabo is the only source to report the use of “Edessa” as an alternate name for Bambyke. Most scholars think Strabo has confused and conflated Bambyke and Edessa (on the 245 Syme (1995) 107. 246 For more on Bambyke/Hierapolis, see Appendix 1.4. 247 A schoinoi is an hour’s travel, see Radt (2002) 8.448; Jansen-Winkeln, “Schoinos”, BNP; Chaumont (1984) 66– 67. Musil calculated the length of Isidore’s schoinoi as approximately 4.7km, (1927) 228. 248 Strabo 16.1.27: “ὑπέρκειται δὲ τοῦ ποταμοῦ σχοίνους τέτταρας διέχουσα ἡ Βαμβύκη, ἣν καὶ Ἔδεσσαν καὶ Ἱερὰν πόλιν καλοῦσιν, ἐν ᾗ τιμῶσι τὴν Συρίαν θεὸν τὴν Ἀταργάτιν.” (Above the river, at a distance of four schoinoi, lies Bambycê, which is also called Edessa and Hierapolis, where the Syrian goddess Atargatis is worshipped.) Lucian De Dea Syria 1,10ff. 76/448 Kallirhoe), 249 but Brodersen suggests that Edessa was an official name of Bambyke. 250 To add to the complication, Strabo’s note about Atargatis could apply to either city . Bambyke was the centre of the Syrian Atargatis cult but that goddess was also worshipped at Edessa into the late empire. 251 Bambyke is quite close to Edessa on the Kallirhoe, so we might expect that the proximity of two towns named Edessa would be needlessly confusing and reduce the likelihood of a Seleucid king applying the name “Edessa” to Bambyke; however the large number of cities named Antioch across the Seleucid realm is a good indication that the kings had no such concern. In particular, if we consider the three important Antiochs in or near the Mesopotamian borderland, Antioch on the Orontes, Antioch on the Euphrates and Mygdonian Antioch, we see that two of them are more commonly associated with other names: Zeugma and Nisibis. Edessa may have been a rarely used ceremonial or honorific name, replaced by , or supplementing, Heirapolis. On the other hand, there is already clear evidence of carelessness in Strabo’s description of the city . While Pliny and Ptolemy correctly locate the site in Cyrrhestice, west of the Euphrates, Strabo places it on different sides of the river in two separate passages: this one describing Mesopotamia in which he seems to place the city to the east of the Euphrates, in fact, in the direction of the famous Edessa, and also in his description of Syria where he places Bambyke west of the Euphrates near Cyrrhestice and calls it a small town (πολίχνη). 252 In his description of Mesopotamia, Strabo may have conflated the descriptions of Bambyke and Edessa which he found in his source for the trade route, placing the conflated description at Edessa but describing Bambyke. 249 Ross (2001) 16; Cohen (2006) 173, 177; Benzinger, RE ‘Bambyke’; Honigmann, RE Suppl. ‘Hierapolis’; Tcherikover, Die hellenistischen stadtegrundungen von Alexander von Grossen bis auf die Romerzeit. Leipzig, 1927, 56; Syme (1995) 107–8. 250 Brodersen (1989) 152, n.2. 251 Ross (2001) 85–90, 109. 252 Strabo 16.2.7: “πρὸς ἕω δ’ ὁ Εὐφράτης ἐστὶ καὶ ἡ Βαμβύκη καὶ ἡ Βέροια καὶ ἡ Ἡράκλεια τῇ Ἀντιοχείᾳ, πολίχνια τυραννούμενά ποτε ὑπὸ Διονυσίου τοῦ Ἡρακλέωνος” (To the east of Antiocheia are the Euphrates, as also Bambyke and Beroea and Heracleia, small towns once ruled by the tyrant Dionysius, the son of Heracleon). Radt (2002) 8.290. Bambyke was a polis with a boule and a demos in the Hellenistic period, Cohen (2006) 173. 77/448 The last community Strabo names is Skenai, which he places 18 schoinoi from Seleucia on the Tigris (διέχουσι δὲ τῆς Σελευκείας αἱ Σκηναὶ σχοίνους ὀκτωκαίδεκα). Strabo calls Skenai a notable city (ἀξιολόγου πόλεως) and locates it on a canal (ἐπί τινος διώρυγος ἱδρυμένης). 253 The name Skenai is not known from any other source, but Sarre and Herzfeld suggest that is a corruption of a city recorded by Stephanus of Byzantium: Maschane (Μασχάνη), a city of the Skenitai Arabs. 254 This important Babylonian city on the Euphrates at the inflow of Naarmalcha canal seems never to have received a consistent transliteration in Greek or Latin and is known variously as Besechana, Macepracta, and, in the Greek text of a Sasanian inscription, Misiche. 255 It is understandable that confusion about the name of this city could occur in the commercial context of Strabo’s description. In Greek, σκηναί are booths from which goods are sold; it would be natural to expect this association when reporting or copying the details of the southern terminus of a merchant route. 256 Conclusion Strabo gives a clear picture of the Romano-Iranian borderland from the perspective of Roman Syria. Although his overall narrative moves from east to west, a predominantly Mediterranean perspective is evident, particularly in the specificity of the information regarding Syria and the proliferation of Greek names and cultural impositions on the landscape of the region. Strabo’s information about οἱ Ἀσσύριοι is more precise in Syria and becomes less so as the narrative progresses further east. Strabo’s approach is fundamentally geographical. Beginning with his denomination of the space 253 Strabo 16.1.27: “διαβάντων δὲ ἡ ὁδός ἐστι διὰ τῆς ἐρήμου ἐπὶ τοὺς τῆς Βαβυλωνίας ὅρους μέχρι Σκηνῶν, ἀξιολόγου πόλεως...“ (...for after they cross the river, the road runs through the desert to Skenai, a noteworthy city situated on a canal towards the borders of Babylonia...”). 254 Sarre and Herzfeld (1911) I.227–28; W eissbach “Σκηναὶ”, RE Series 2, 5.470. Steph. Byz. s.v . Μασχάνη, “πόλις πρὸς τῶν Σκηνιτῶν Ἀράβων. Κουάδρατος ὀγδόῳ Παρθικῶν. τὸ ἐθνικὸν Μασχανεύς” (Meineke 437). 255 Isidore (1) calls it Besechana and places it 31 schoinoi from the Naarmalcha canal. In his commemorative inscription at Naqš-i Rustam, Shapur calls it Misiche: ŠKZ (Res Gestae Divi Saporis) 4-5, 9-10 (Greek text with French translation: Maricq (1958); the main edition is Huyse (1999)); Gawlikowski (2007) 131. Ammianus calls it Macepracta: Amm. Marc. 24.2.6. 256 This route is discussed in detail in Chapter 5. 78/448 itself, Mesopotamia is defined by the relationship of the space to the two rivers which enclose it. Strabo delimits the boundaries of Mesopotamia primarily according to geographical features: the Euphrates and Tigris rivers and the Taurus Mountains. Furthermore, when he divides the space into smaller units, be locates those units by reference to geographical features: Mygdonia is near the mountains and the Euphrates (ἡ παρόρειος), Gordyene is near the Tigris (Πρὸς τῷ Τίγρει), and the southern parts are defined by their relationship to the desert. Although most of the sites in Mesopotamia are not topologically located, those that are defined in this way are defined by their topological relationship to geographic features: Nisibis is under Mount Masius (ὑπὸ τῷ Μασίῳ ὄρει) and Anthemusia, Bambyke and Skenai are all located relative to nearby rivers or canals. These topological relationships also give the work a greater sense of two- dimensionality . Pliny the Elder In his first reference to Mesopotamia in book five of his Natural History , Pliny’s stated definition of Mesopotamia is the same as Strabo’s: “eadem Mesopotamia inter Euphraten et Tigrin”. 257 However, rather than addressing Mesopotamia in a single discussion, Pliny divides his description of that geographical region between the two rivers into two well-separated sections: the first comes in book five as part of his description of Syria and the second in book six with his account of the rest of Asia. In making this narrative division, Pliny represents the space along ideological lines rather than strictly according to geographical divisions. These ideological implications will be explored in Chapter Four; in the present chapter, their existence is 257 Pliny NH 5.66; see also 6.25, “Mesopotamiae... inter duos amnes sitae”. Strabo 16.1.21: εἴρηται δ’ ὅτι κεῖται τοῦ Εὐφράτου μεταξὺ καὶ τοῦ Τίγριος καὶ διότι ὁ μὲν Τίγρις τὸ ἑωθινὸν αὐτῆς μόνον κλύζει πλευρόν, τὸ δ’ ἑσπέριον καὶ νότιον ὁ Εὐφράτης (As I have said, it lies between the Euphrates and the Tigris; and the Tigris washes its eastern side only , whereas the Euphrates washes its western and southern sides). Quoted in full above, p.44. This was also the method of Arrian and perhaps Nearchos, BNJ 133 F 1 XVI = Arr. Indica 42.3: “ὃς ῥέων ἐξ ᾽Αρμενίης παρὰ πόλιν Νῖνον, πάλαι ποτὲ μεγάλην καὶ εὐδαίμονα, τὴν μέσην ἑωυτοῦ τε καὶ τοῦ Εὐφράτου ποταμοῦ <χώρ>ην Μεσοποταμίην ἐπὶ τῶιδε κληίζεσθαι ποιέει.” (This river flows from Armenia, past the city of Ninos, which of old was once great and prosperous, and causes the land between itself and the river Euphrates to be called Mesopotamia for that reason.) Trans. Whitby , ‘Nearchos (133)’, BNJ. 79/448 important for what it reveals about Pliny’s priorities as a geographical writer. Namely that geographic accuracy was not his primary concern. Pliny used geographical categories and features to frame and locate certain parts if his narrative, but most of the topological relationships between geographical features of the Mesopotamian borderland are linear. This stands in contrast with Strabo whose understanding of planar topology is evident in his geographical descriptions. Book Five: Syria Part of Pliny’s description of Mesopotamia is included in his account of Syria, the easternmost province of the Roman empire at Pliny’s time. After describing Egypt up to Pelusium, he pauses to describe Syria in general terms (see Map 9: Pliny's Syrian Overview). [66] Iuxta Syria litus occupat, quondam terrarum maxuma et plurimis distincta nominibus. namque Palaestine vocabatur qua contingit Arabas, et Iudaea et Coele, dein Phoenice et qua recedit intus Damascena, ac magis etiamnum meridiana Babylonia, eadem Mesopotamia inter Euphraten et Tigrin quaque transit Taurum Sophene, citra vero eam Commagene et ultra Armeniam Adiabene, Assyria ante dicta, et ubi Ciliciam attingit Antiochia. [67] longitudo eius inter Ciliciam et Arabiam CCCCLXX [m.]p.. est, latitudo a Seleucia Pieria ad oppidum in Euphrate Zeugma CLXXV [m.p.]. qui subtilius dividunt, circumfundi Syria Phoenicen volunt et esse oram maritimam Syriae, cuius pars sit Idumaea et Iudaea, dein Phoenicen, dein Syriam. id quod praeiacet mare totum Phoenicium appellatur. ipsa gens Phoenicum in magna gloria litterarum inventionis et siderum navaliumque ac bellicarum artium. 258 This description shows the geographic ambiguity of the term “Syria” which for Pliny encompasses the entire fertile crescent including Mesopotamia, Babylonia and Adiabene. 259 Pliny indicates the contemporary 258 Pliny , NH, 5.66-67: “Next along the coast is Syria, at one time the greatest of lands and divided with many different names. For it was called Palestine where it touches the Arabians, and Judaea and Coele, then Phoenicia and further inland, Damascena, and then even further, southern Babylonia. It is called Mesopotamia between the Euphrates and Tigris rivers, Sophene across the Taurus. Then, Commagene on this side of Sophene; Adiabene, previously called Assyria, on the other side of Armenia and Antiochia where it touches Cilicia. [67] It is 470 m.p in length between Cilicia and Arabia, and 175 m.p. wide from Seleucia Pieria and to Zeugma, a town on the Euphrates. Those who divide it into smaller parts want Phoenecia to be surrounded by Syria and the coast to be maritime Syria, of which Idumaea and Judaea are a part, then Phoenicia, then Syria. The whole sea which lies offshore they call the Phoenecian Sea. The entire race of Phoenecians have great fame from the discovery of letters and the constellations and the military and naval arts.” 259 The broad application can also be seen in Strabo 16.1.1-2. 80/448 denomination of Syria by his use of present tense verbs to express topological relationships throughout the passage, including occupat to describe Syria broadly . However, most of the names that Pliny lists are governed by vocabatur in the imperfect. By using the incomplete past tense, Pliny indicates a continuing state in the past during which the areas he names had separate and individual denominations now superseded by their aggregation into Roman Syria. In this way his text also defines the space of Syria in historical terms. Moreover, these other methods of dividing the space into smaller regions clearly illustrate that the space labelled “Syria” was available for construction and redefinition according to the desires and requirements of the author. In this passage, Pliny gives verbal and geometric definitions for Syria; in the narrative that follows, he seldom strays beyond those geometric dimensions: between Antioch and Zeugma (a Seleucia 81/448 Map 9: Pliny's Syrian Overview Pieria ad oppidum in Euphrate Zeugma) and Cilicia and Arabia (inter Ciliciam et Arabiam). 260 These dimensions clearly do not correspond to Pliny’s own broader definition of “Syria” which includes all of Mesopotamia, Babylonia and Adiabene, rather, they correspond more or less to the limits of the Roman province of Pliny’s time. Pliny structures his description of Syria according to lists. The main line of the narrative follows the Syrian coast north from Egypt then the Euphrates south from the Taurus. He diverges from and returns to this structure repeatedly , including a notable section in which he includes two lists of people and places in North Syria detached from the topography of the region. 261 Pliny disguises the alphabetical and atopological nature of these lists well, but his “Coele Syria” and “reliqua Syria” do not cover distinct geographical areas (Map 10: Pliny's Syria). I will discuss these lists and their relationship both to Pliny’s narrative of the region 260 These dimensions are reasonably accurate. Depending on which start and end points one selects, the straight line distance between Cilicia and Arabia is 600-700 km; 470m.p. is 696 km. The straight line distance between Seleucia Pieria (Antioch’s port, close to the mouth of the Orontes) and Zeugma is a little over 200km; 175m.p.is 259 km. 261 Pliny NH 5.81-82: “[81] Nunc interiora dicantur. Coele habet Apameam, Marsya amne divisam a Nazerinorum tetrarchia, Bambyken, quae alio nomine Hierapolis vocatur, Syris vero Mabog — ibi prodigiosa Atargatis, Graecis autem Derceto dicta, colitur —, Chalcidem cognominatam Ad Belum, unde regio Chalcidena fertilissima Syriae, et inde Cyrresticae Cyrrum, Gazetas, Gindarenos, Gabenos, tetrarchias duas quae Granucomatitae vocantur, Hemesenos, Hylatas, Ituraeorum gentem et qui ex his Baethaemi vocantur, Mariamnitanos, [82] tetrarchiam quae Mammisea appellatur, Paradisum, Pagras, Penelenitas, Seleucias praeter iam dictam duas, quae Ad Euphraten et quae Ad Belum vocantur, Tardytenses. reliqua autem Syria habet, exceptis quae cum Euphrate dicentur, Arethusios, Beroeenses, Epiphanenses ad Orontem, Laodicenos, qui Ad Libanum cognominantur, Leucadios, Larisaeos, praeter tetrarchias in regna discriptas barbaris nominibus XVII.” (Let the interior now be described. Coele Syria has Apamea, separated from the tetrarchy of the Nazerini by the river Maysa; Bambyke (which his also called Hierapolis, and in fact Mabog by the Syrians – here the monstrous Atargatis is worshipped, called Derceto by the Greeks) Chalcis, named On the Belus, from which the very fertile region of Syria, Chalcidene, is named, and then Cyrrus of Cyrrestica, the Gazetae, Gindareni, Gabeni and the two tetrarchies called Granucomatitae, the Hemeseni, Hylatae, the tribe of Ituraei (and one branch called the Baethaemi), the Mariamnitani, [82] the tetrarchy called Mammisea, Paradisum, Pagrae, Penelenitae, two Seleucias (beyond that already mentioned), one called “on the Euphrates” the other “on the Belus”, and the Tardytenses. The rest of Syria, except for the parts to be addressed with the Euphrates, has the Arethusii, the Beroeenses, the Epiphanenses on the Orontes, the Laodiceans (called “on Lebanon”), the Leucadii, the Larisaeos, and seventeen more tetrarchies with barbarian names divided into kingdoms.) 82/448 and to his overall project in Chapter 4. After describing the inland parts of Northern Syria, Pliny transitions to a description of the course of the Euphrates. He begins at its source in Armenia and proceeds downstream to the marshes of Babylonia, with a short digression to present Palmyra. It is in the course of this description that Pliny treats several cities of reliqua Syria that he previously excepted from that list. 262 I will discuss Pliny’s narrative under four sections, describing in turn the delimitations, denominations and topological relationships of Northwest Mesopotamia, Commagene, the cities and people along the Euphrates, and Palmyra. 262 Pliny NH 5.82: “reliqua autem Syria habet, exceptis quae cum Euphrate dicentur”, translated above, p.49. 83/448 Map 10: Pliny's Syria Northwest Mesopotamia I conceptualise Pliny’s description of the Mesopotamian borderland in book five as divided into four parts. For convenience, I will refer to the first of these parts as northwest Mesopotamia for reasons which I explain shortly , but it should be noted that the analytical category is mine, not Pliny’s. Pliny’s description of this area refers to three entities: Orroeon (Osrhoene), the praefectura Mesopotamiae, and the Praetavi Arabs. What entities these terms reflect is unclear, but while they seem to include the entire fertile shelf of northern Mesopotamia, the topological clues which Pliny provides show that his focus is on the western part of northern Mesopotamia, west of the Balikh river, thus, northwest Mesopotamia. Pliny begins his narrative itinerary along the Euphrates with a digression inland to describe the regional topology of northwest Mesopotamia: Arabiam inde laeva, Orroeon dictam regionem, trischoena mensura, dextraque Commagenen disterminat, pontis tamen, etiam ubi Taurum expugnat, patiens. 263 “Orroeon” transliterates a Greek genitive plural which in turn transliterates the Syriac name for Edessa, ܝܗܪܘܐ (’WRHY , Orhai), thus the space commonly referred to as Osrhoene. 264 In the geographical context of the downstream journey on the Euphrates, this description is quite accurate: Osrhoene (one of several regions called Arabia in antiquity) 265 is to the left of the Euphrates, and Commagene is to the right. The implication of the text is that the two regions faced each other across the river, as they did in reality . However the narrative context of the passage undercuts the geographical accuracy of this description. This passage seems to be interposed amidst a description of the course of the Euphrates through the Taurus. Preceding this passage, Pliny had described the course of the Euphrates through Armenia; here he appears to begin a description of the Euphrates south of the Taurus, but suddenly skips back through the range to describe the 263 Pliny NH 5.85: “By a distance of by three schoinoi, the Euphrates separates the region of Arabia called Osrhoene on the left and Commagene on the right, and allows a bridge, even where it smashes out of the Taurus.” 264 Millar (1993) 456. Osrhoene will be discussed in more detail in Chapter 2. 265 Shahid (1984) 3–5, 7–9. 84/448 eastward turn of the Euphrates at Claudiopolis in Cappadocia and the violent passage of the river through the mountains. 266 The poetic language the following passage and the casual rearrangement of the spaces’ geographical order serves as a warning that building an accurate topological image of the world was not Pliny’s primary concern. 267 The essential structure of Pliny’s description of Mesopotamia is a simple set of lists largely unconcerned with topological specificity at the local level. 268 Having introduced “Arabia” (Osrhoene) Pliny diverges from the river to describe the regions of Mesopotamia: Arabia supra dicta habet oppida Edessam, quae quondam Antiochia dicebatur, Callirhoem, a fonte nominatam, Carrhas, Crassi clade nobile. iungitur praefectura Mesopotamiae, ab Assyriis originem trahens, in qua Anthemusia et Nicephorium oppida. mox Arabes qui Praetavi vocantur; horum caput Singara. 269 Pliny’s description works on two scales: at the broader scale three categories of space are placed in topological relation to each other: the praefectura Mesopotamia is joined (iungitur) to Arabia (Osrhoene), and thereafter (in the form of a list, using mox) come the Praetavi Arabs (Arabes qui Praetavi vocantur). It is unlikely that this list reflects an intra-urban itinerary across northern Mesopotamia; a route passing in order through Edessa, Carrhae, Anthemusia, Nikephorion, and Singara would have been a very circuitous route (see Map 266 Pliny NH 5.85: “apud Claudiopolim Cappadociae cursum ad occasum solis agit. primo hunc illic in pugna Taurus aufert victusque et abscisus sibimet alio modo vincit ac fractum expellit in meridiem. ita naturae dimicatio illa aequatur, hoc eunte quo vult, illo prohibente ire qua velit.” (At Claudiopolis in Cappadocia [the Euphrates] turns its course to the west. There, the Taurus redirects the river for the first time in this conflict and despite being conquered and cut in half it wins in another way and expels it, broken, to the south. Thus this struggle of nature is a draw , the river breaking through as it wished, but the mountain stopping its desired course). 267 This passage is discussed further in Chapter 4. 268 In the introduction to his geographical section Pliny (NH 3.1-2) describes how he will go about constructing his geography: “Locorum nuda nomina et quanta dabitur brevitate ponentur, claritate causisque dilatis in suas partes; nunc enim sermo de toto est.” (The bare names of places will be set down, and with the greatest brevity available, their celebrity and its reasons being deferred to the proper sections; for my topic now is the world as a whole.) Pliny’s project is discussed in detail in Chapter 4. 269 Pliny NH 5.86: “Arabia, above mentioned, has the towns of Edessa, formerly called Antiochia; Callirhoë, named from its fountain, and Carrhæ, famous for the defeat of Crassus. Adjoining to this is the praefecture of Mesopotamia, which derives its origin from the Assyrians, and in which are the towns of Anthemusia and Nicephorium. Next are the Arabians, called Praetavi whose capital is Singara.” 85/448 11: Mesopotamian Sites in Pliny NH 5). At the closer scale, within those topologically specified regions, Pliny lists the important settlements of those spaces without relational cues. The narrative structure is well disguised with descriptive asides, but is fundamentally a series of lists of people and places organised by regions: Osrhoene contains the oppida Edessa, Callirhoe, and Carrhae; Mesopotamia contains Anthemusia and Nicephorium; the land of the Praetavi Arabs contains Singara. This pattern of topologically located regions with listed contents is evident throughout Pliny’s geographical narrative. 270 One of the most difficult toponyms in Pliny’s account of the borderland is the term “praefectura Mesopotamiae”. That designation falls in a list between Osrhoene and the Praetavi Arabs around Singara and, like the other elements of the list, refers to a spatial area containing urban areas: “iungitur praefectura Mesopotamiae, ab Assyriis originem trahens, in qua Anthemusia et Nicephorium oppida.” The progression from Osrhoene to praefectura Mesopotamiae to the region of Singara brings to mind the later Roman provincial division, implemented perhaps by Trajan and certainly by Lucius Verus, of the region between Osrhoena, a province centred on Edessa, and Mesopotamia, a province with its capital at Amida on the Tigris to the east. 271 However, the two oppida mentioned as part of the praefectura Mesopotamiae, Anthemusia and Nicephorium, lie west and south of Osrhoene as Pliny defines it (that is, around Edessa and Carrhae) as Map 11 shows. Anthemusia and Nicephorium are usually depicted as part of Roman Osrhoena rather than Roman Mesopotamia. 270 Some examples from his description of Syria:the coastal cities of Palestina are described as a periplus and interior cities as a list (5.68-70); Pliny specifies the location of the Decapolis within Syria and relative to Judaea, then describes the polities within it as lists (5.74); the coast of Phoenecia is described as a periplus with a brief topological aside to describe the relationship between the Lebanon and Anti-Lebanon ranges, Phoenecia, Palestina and the Decapolis (5.75-8). Examples from the rest of his geographical books abound, see, for example, his passages on Spain (NH 3.6-30). 271 See Chapter 2. Amida is modern Diyarbakır in Turkey. It had also been a provincial capital in the 9th-7 th century Neo-Assyrian empire, Postgate (1995). 86/448 Pliny’s note that the origin of the “praefectura Mesopotamiae” is derived “ab Assyriis” raises the question of what Pliny meant by ab Assyriis and whether we can infer anything about the geographical space from that phrase. Did Pliny mean to indicate a degree of continuity between a Parthian political structure and that employed by the Assyrians, even if the extent of that continuity was toponymic or geographical? In this phrase, “Assyrians” refers to a collective, but is not specific as to whether it refers to a political entity (the Assyrian empire) or an cultural entity (the Assyrian people), each of which varied in geographical extent over time. As a political term, the geographical bounds of Assyria varied from the 87/448 Map 11: Mesopotamian Sites in Pliny NH 5 original heartland around Aššur to an empire encompassing much of the fertile crescent, from southern Babylonia to southern Palestine and into the Taurus mountains. 272 Pliny is certainly aware of the former use of “Assyria”; in his initial description of Syria, he notes that Adiabene was once called Assyria (“Adiabene Assyria ante dicta”). 273 The terms Syria and Assyria were used somewhat interchangeably by classical authors. 274 Using a recently published Hieroglyphic Luwian and Phoenician bilingual inscription, Rollinger has demonstrated that the two terms were also used interchangeably in 8 th century BCE Anatolia and were most likely encountered by the Greeks as synonyms already . 275 Numerous Greek and Roman authors mention and discuss the Assyrians. The mythologised figure of the Assyrian queen Semiramis held a particular fascination. Diodorus, Pompeius Trogus and Strabo all describe her and her works and deeds at some length. 276 Moreover, Strabo refers to “those who have written histories of the Syrian Empire” (οἱ δʼ ἱστοροῦντες την Συρων ἀρχην). 277 The historical memory of the Assyrians may have been confused with legend to a degree by the Roman period, but it was nevertheless present, available for use by Roman authors, and still attached to the geographical space of the historical neo- Assyrian empire. Frye argues that “Assyrian” and “Syrian” were ethnolinguistic terms by which the Greeks understood the Aramaic-speaking population of the fertile crescent. 278 Parpola goes further to propose that that Aramaic-speaking population had a sense of national identity as “Assyrians”. 279 Nöldeke and Schwartz see the use of “Assyria” as referring to the political formation of the neo-Assyrians and the geographical 272 Wilkinson et al. (2005) 46, fig. 15; Parpola et al. (2001). 273 Pliny NH 5.66, quoted above, p.43. 274 Rollinger (2006). Hdt 7.63. Frye (1992) 283; contra Parpola (2004) 21; and less convincingly Joseph (1997). 275 Rollinger supports the main thesis of Frye (1992); Parpola (2004) 16–18 also notes that unstressed initial vowels or syllables were often dropped in neo-Assyrian. 276 For example: Diodorus 2.5ff.; Justin 1.1-3; Strabo 2.1.16; 2.1.31; 15.1.5-6; 15.2.5; 16.1.2. See Chapter 4 for a full discussion. 277 Strabo 16.1.2. 278 Frye (1992) 282; Heinrichs (1993) 104. 279 Parpola (2004) 8–11, although Parpola’s use of the United States to illustrate his argument that new citizens will attain a homogeneous national identity by the third generation unintentionally problematises his position on the presence of an Assyrian national identity in neo-Assyrian empire. Heinrichs (1993) discusses the linguistic and cultural vectors by which the term “Assyrian” came to be adopted by the modern Assyrians. 88/448 scope of that empire. 280 Although the terms used by these scholars are controversial, the evidence they present speaks to a high level of cultural coherence among the Aramaic speaking population of Pliny’s Mesopotamian borderland. 281 In writing of a praefectura, a technical term of administrative organisation, Pliny most likely means Assyria as a political unit. 282 The neo-Assyrian empire was known to the Greeks from the earliest complete historical source, Herodotus. The Ionian historian promises, but does not deliver, a history of Assyria. 283 However, his account of the rise of the Achaemenid empire begins with the destruction of the Assyrian empire. In that narrative, he conceives of Assyria as the wider political formation of the neo-Assyrian empire rather than the narrower geographical area around the Assyrian capitals on the Tigris. 284 Specifically , Herodotus considers Babylonia (Akkad) as an integral part of Assyria, as does Strabo. 285 Later, Herodotus 280 Nöldeke (1881); Schwartz (1931); Schwartz (1932); Rollinger (2006). 281 McGing (2012). 282 Pliny most often uses praefectura to refer to subdivisions of Roman administrative authority , usually defined spatially. He specifically refers to the nomes of Egypt as praefecturae (Pliny NH 5.49; 19.11; 28.121; 36.87) as well as the strategiae of Armenia (Pliny NH 6.27; 5.83). On other occasions, he uses praefectura to refer to a community in Etruria (Pliny NH 3.52), the office of the Urban Prefect at Rome (Pliny NH 7.62), and a personal charge given to a slave (Pliny NH 28.56). 283 Hdt. 1.184: “Τῆς δὲ Βαβυλῶνος ταύτης πολλοὶ μέν κου καὶ ἄλλοι ἐγένοντο βασιλέες, τῶν ἐν τοῖσι Ἀσσυρίοισι λόγοισι μνήμην ποιήσομαι, οἳ τὰ τείχεά τε ἐπεκόσμησαν καὶ τὰ ἱρά, ἐν δὲ δὴ καὶ γυναῖκες δύο.” (Now among the many rulers of this city of Babylon (of whom I shall make mention in my Assyrian History), who finished the building of the walls and temples, there were two who were women.) Loeb trans. 284 Hdt 1.188-200. When describing the important spaces of the Persian empire, Herodotus (4.39) specifies Persia, Assyria and Arabia as the major parts: “Ἡ δὲ δὴ ἑτέρη ἀπὸ Περσέων ἀρξαμένη παρατέταται ἐς τὴν Ἐρυθρὴν θάλασσαν, ἥ τε Περσικὴ καὶ ἀπὸ ταύτης ἐκδεκομένη ἡ Ἀσσυρίη καὶ ἀπὸ Ἀσσυρίης ἡ Ἀραβίη· λήγει δὲ αὕτη, οὐ λήγουσα εἰ μὴ νόμῳ, ἐς τὸν κόλπον τὸν Ἀράβιον, ἐς τὸν Δαρεῖος ἐκ τοῦ Νείλου διώρυχα ἐσήγαγε· μέχρι μέν νυν Φοινίκης ἀπὸ Περσέων χῶρος πλατὺς καὶ πολλός ἐστι, τὸ δ’ ἀπὸ Φοινίκης παρήκει διὰ τῆσδε τῆς θαλάσσης ἡ ἀκτὴ αὕτη παρά τε Συρίην τὴν Παλαιστίνην καὶ Αἴγυπτον, ἐς τὴν τελευτᾷ· ἐν τῇ ἔθνεά ἐστι τρία μοῦνα.” (But the second, beginning with Persia, stretches to the Red Sea, being the Persian land, and next the neighbouring country of Assyria, and after Assyria, Arabia; this promontory ends (yet not truly but only by common consent) at the Arabian gulf, whereunto Darius brought a canal from the Nile. Now from the Persian country to Phoenice there is a wide and great tract of land; and from Phoenice this promontory runs beside our sea by the way of the Syrian Palestine and Egypt, which is at the end of it; in this promontory there are but three nations.) Loeb trans. It would be remarkable if lower Mesopotamia failed to be included in such a list, but it is consistent with the rest of his work for Herodotus to have subsumed Babylonia into “Assyria”. 285 Strabo 16.1.1, discussed above, p.41. On Babylonia becoming a de facto part of Assyria in 731 BCE when Tiglath-Pileser III, assumed the kingship of Babylon, see Parpola (2004) 8, n.17. There are a number of correspondences between Herodotus’ description of Babylonia and Strabo’s, not least of all that both include a lengthy description of sacred prostitution in Babylon: Herodotus, 1.199; Strabo 16.1.20. 89/448 equates Syria and Assyria and defines the latter as the lands inhabited by the Assyrians. 286 Pliny’s note on the Assyrian origins of the praefectura Mesopotamiae makes sense in the context of the neo-Assyrian administration. Both Nisibis and Harran (Carrhae) were important regional administrative centres of the neo-Assyrian empire from the ninth to seventh centuries. 287 Nisibis is notably absent from Pliny’s description, but would be a natural centre to describe a political division spanning Mesopotamia and with Assyrian administrative heritage. Assyrian administrative divisions in the fertile crescent were not significantly altered by the subsequent empires and dynasties. The Achaemenids and Seleucids seem to have retained much the same organisational structure in the region. 288 In Pliny’s time, all these regions to the east of the Euphrates were nominally Parthian territory: “nominally”, in that the Euphrates formed a conceptual and symbolic boundary in geopolitical terms and was reinforced as such at the highest levels of government by repeated performance of inter-state rituals on the river during the first century CE. 289 Pliny never explicitly mentions this division of space, and as we shall see, deliberately complicates it. 290 Nevertheless, Pliny’s use of praefectura here, probably indicates his understanding of Mesopotamia as a unit of Parthian administration. 291 How accurately his understanding reflects that administrative reality is difficult to determine. According to Pliny’s understanding of northern Mesopotamia, Anthemusia and Nicephorium both lay within this Parthian praefectura, while Carrhae and Edessa were associated with the Orroeni Arabs. 292 286 Hdt. 2.17: “...Αἴγυπτον μὲν πᾶσαν εἶναι ταύτην τὴν ὑπ’ Αἰγυπτίων οἰκεομένην, κατά περ Κιλικίην τὴν ὑπὸ Κιλίκων καὶ Ἀσσυρίην τὴν ὑπὸ Ἀσσυρίων...” (...Egypt is all that country which is inhabited by Egyptians, even as Cilicia and Assyria are the countries inhabited by Cilicians and Assyrians severally...) Loeb trans; 7.63: “Οὗτοι δὲ ὑπὸ μὲν Ἑλλήνων ἐκαλέοντο Σύριοι, ὑπὸ δὲ τῶν βαρβάρων Ἀσσύριοι ἐκλήθησαν.” (These are called by Greeks Syrians, but the foreigners called them Assyrians.) Loeb trans. 287 Nisibis, above p.44. Harran: Appendix 1.6. 288 See above, pp.43f. 289 For example: Plut. Sulla 5.3-4; Luc. 24; Vell. Pat. 2.101.1-3; Dio 55.10.18-19; Tac. Ann. 2.58.1-2. For a discussion of the notion of “independence” within the “Parthian Commonwealth”, see de Jong (2013) 148–51. 290 See Chapter 4. 291 Pliny most often uses praefectura to translate a non-Roman administrative term, see n.168, p.54. 292 Nikephorion and Anthemusia, see Appendices 1.13 and 1.3 respectively. 90/448 The final division in this part of Pliny’s narrative are the Arabes qui Praetavi vocantur with their lands around Singara. 293 The Praetavi are otherwise unknown. Retsö suggests that the -tav- element of their name may indicate a relationship to taienoi which became a general term for people of the Syrian desert and close to Adiabene in the third century CE. 294 Thus Pliny’s description of the trans-Euphrates area in book five presents the space as containing two Arab ethnic units centred on cities, and a praefectura with no ethic description. The territory of the Orroeni around Edessa and the Praetavi around Singara lie within a region broadly administered in some relationship to a Parthian administration (praefectura) of Mesopotamia as a whole. In Pliny’s time, the Abgarid dynasty of Edessa was a political entity within the Parthian system, but is that what Pliny is describing here? The space Pliny calls Orroenian Arabia lay on the east bank of the Euphrates. It could be described as opposite Commagene and contained the oppida of Edessa and Carrhae. This last note, the assignation of Carrhae to the territory of Orroenian Arabia, has been disputed. Ross questions Pliny’s value for establishing the limits of Osrhoene not only on the basis of the inclusion of Carrhae in book five, but on the description of the Tigris in book six as separating the Orroei from the Adiabeni. 295 It is extremely unlikely that the territory of the Abgarid kings extended to the Tigris opposite Adiabene. Such a swathe would encompass the territories of several other northern Mesopotamian cities including Nisibis, Rhesaina and Singara. Ross concludes that Pliny has confused the inhabitants of Osrhoene with the other sedentary Arab peoples of the region. However, this may be a misunderstanding of Pliny’s text. It is begging the question to ask if Pliny is 293 For Singara, see Appendix 1.17. 294 Retso (2003) 415, 521. Pliny does not mention Hatra, but it may have been been within the sphere of the Praetavi, Retso (2003) 440. 295 Ross (2001) 22–23. Pliny , NH 5.86; 6.129: “Tigris autem ex Armenia acceptis fluminibus claris Parthenia ac Nicephorione Arabas Orroeos Adiabenosque disterminans et quam diximus Mesopotamiam faciens....” (However, once the notable rivers flowing from Armenia, the Parthenias and Nicephorion, have been received, the Tigris divides the Orroei Arabs from the Adiabeni Arabs and creates Mesopotamia, which we mentioned above...). 91/448 correct to assign Carrhae to the territory of the Orroeni when Pliny’s description does not attribute a political status to “Arabiam inde laeva, Orroeon dictam regionem”. Pliny uses no political terminology in these descriptions. Just as Strabo refers broadly to all of the settled plain of northern Mesopotamia as the land of the Mygdonian Arabs, so Pliny attributes a broad area of northern Mesopotamia to Orroenian Arabs. When Pliny describes the extent of Greater Armenia in book six, he refers to the Tigris and Euphrates rivers as delimiting Mesopotamia and the Arabes Orroei as occupying it: “[Mesopotamia] interest ibi tenent Arabes Orroei”. 296 Pliny may well have confused the inhabitants of Osrhoene with the other sedentary Arab peoples of the region as Ross argues, but his mistake is not necessarily one of misattributing territory to the Abgarid dynasty of Edessa, but of using an ethnic term to apply to a collection of neighbouring units which might more accurately be described by different ethnic terms, as Strabo may have done in his description of Mygdonia and the Mygdones. If this is the case, the different name in the two authors may show a shift in priority from a view of Mesopotamia centred on Nisibis to one centred on Edessa, that is, from the broader region of Mesopotamia generally to the immediate border region of Osrhoene on the banks of the Euphrates). 297 This perspective would fit with the recent annexation of Commagene which pushed Roman direct control as far as the Euphrates and made Osrhoene an immediate neighbour of Roman space. Pliny certainly focuses on the western edge of the Mesopotamian borderland in this section; Singara is the only site he mentions that lies east of the Balikh river (see Map 11: Mesopotamian Sites in Pliny NH 5). Commagene After dealing with Mesopotamia in this way , Pliny’s description of Commagene immediately 296 Pliny NH 6.25: “Armenia autem Maior incipit a Parihedris montibus, Euphrate amne, ut dictum est, aufertur Cappadociae et, qua discedit Euphrates, Mesopotamiae haut minus claro amne Tigri. Utrumque fundit ipsa, et initium Mesopotamiae facit inter duos amnes sitae; quod interest ibi tenent Arabes Orroei.” (Greater Armenia begins at the Parihedri Mountains, and is separated from Cappadocia, as we have said, by the Euphrates and, when the Euphrates turns aside, from Mesopotamia by the equally famous river Tigris. Both rivers rise in Armenia, and it forms the beginning of Mesopotamia, the tract of country lying between those two rivers; the intervening space is occupied by the Orroean Arabs.) Loeb trans. 297 Osrhoene’s position as a small kingdom in an inter-imperial borderland will be discussed in Chapter 2. 92/448 following is relatively specific. This is the second part of his description of the Mesopotamian borderland. He jumps abruptly from his sentence on the Praetavi Arabs back to the Euphrates at Samosata: a Samosatis autem latere Syriae Marsyas amnis influit. Cingilla Commagenen finit, Imeneorum civitas incipit. oppida adluuntur Epiphania et Antiochia, quae Ad Euphraten vocantur, item Zeugma LXXII [m.]p. a Samosatis, transitu Euphratis nobile. ex adverso Apameam Seleucus, idem utriusque conditor, ponte iunxerat. [87] qui cohaerent Mesopotamiae, Rhoali vocantur. at in Syria oppida Europum, Thapsacum quondam, nunc Amphipolis; Arabes Scenitae. ita fertur usque Suram locum, in quo conversus ad orientem relinquit Syriae Palmyrenas solitudines, quae usque ad Petram urbem et regionem Arabiae Felicis appellatae pertinent. 298 Pliny identifies the border of Commagene, along with other oppida along the river, including the twin-cities of Zeugma (Seleucia on the Euphrates) and Apamea, which he locates relative to the capital (Samosata and Zeugma can be seen on Map 4: The Geographical Boundaries of Strabo's Mesopotamia, the post- Commagene sites on Map 12: Pliny's Euphrates Itinerary, the rest of the sites are unknown). However, this is something of a literary rearrangement of a Euphrates periplus rather than a truly topological description of Commagene. 299 Supported by the grammar and the content, Pliny’s narrative follows the river south from the Taurus. Grammatically , the passage comprises a list governed by verbs of connection (influit, finit, incipit, iunxerat, fertur usque), the main form of conjunctive grammar used. 300 The places described by Pliny are mostly points on the Euphrates comprising either a site (cities and towns) or the intersection of routes, rivers and territorial boundaries. Sites include the inflow of the Marsyas amnis and the oppida Epiphania, Antiochia, Zeugma (also described as an intersection, transitu Euphratis nobile) and Europus-Thapsacus- 298 Pliny NH 5.86-87: “Below Samosata, on the side of Syria, the river Marsyas flows into the Euphrates. Cingilla ends the territory of Commagene, and the city of the Immei begins. The towns Epiphania and Antiochia are washed by the river and are both known as Epiphania and Antiochia on the Euphrates; also Zeugma, seventy- two miles from Samosata, is famous for the passage there across the Euphrates. Seleucus, the founder of both cities, united Zeugma and Apamea on the opposite bank by a bridge. [87] The people on the edge of Mesopotamia are called Rhoali. In Syria is the town Europus, once Thapsacus, now Amphipolis; and Skenitai Arabs. The river flows to a place called Sura where it turns to the east and leaves the Palmyrene desert of Syria. That desert continues as far as the city of Petra and the region of Arabia called Felix.” 299 The same could be said of the specificity of the places mentioned in Pliny’s description of the Jordan valley (5.71-73). 300 Stylistically , Pliny connects the sentences and clauses of the passage sparsely with those verbal connections and little else. At times the passage becomes a simple list. 93/448 Amphipolis. Intersections include the crossing at Zeugma and the boundary of the Imeneorum civitas on the Euphrates. Except for the boundary of Imeneorum civitas the topology of area features is usually very indistinctly described, in particular those areas inhabited by the Rhoali and Arab Skenitai. Pliny’s riverine periplus continues south along the Euphrates, with a digression on the desert cities. Beyond locating it in the region, Pliny devotes little attention to Commagene in his narrative. He places the kingdom within Syria in his initial description of the latter space, then on the Euphrates opposite Osrhoene on his description of the river. 301 In his description of inland north Syria, he mentions none of the Commagenean cities. 302 The passage above is the most detailed: “a Samosatis autem latere Syriae Marsyas amnis influit. Cingilla Commagenen finit, Imeneorum civitas incipit.” Aside from the reference to the Commagenean capital Samosata, Pliny’s account is quite obscure. 303 The Marsyas river has been identified as the Merzumen Su, a major western tributary of the Euphrates that enters that river at Rumkale between Zeugma and Samosata (Rumkale is marked on Map 2); Pliny is the only author to mention it. 304 Pliny is also the only author to mention Cingilla, which he places at the limit of Commagene. Given that Pliny is describing the course of the Euphrates from its source, the logic of his narrative should place Cingilla on the Euphrates at the southern boundary of Commagene. 305 Pliny is likewise the only ancient source to mention the civitas Imeneorum. 306 Pliny is usually clear to mention tribal units that he considers Arabs, whether more pastoralist (Skenitai) or settled (as the Praetavi and Orroeni seem to be). 307 This civitas is probably to be 301 In Syria: Pliny NH 5.66, quoted above, p.49. Opposite Osrhoene: Pliny NH 5.85. 302 Mentioned above, p.43. 303 On Samosata, see below , p.43. 304 Kennedy (1998c) 148, 151; Comfort (2000) 117. Not to be confused with Massyas, the Bekaa valley in Lebanon. Xenophon mentions a river called Maskas (“τὸν Μάσκαν ποταμόν,” Anab. 1.5.4), but this is further south, in the desert region between Thapsacus and Babylonia. 305 However, note that Pliny has shown his willingness to alter his topological sequence for rhetorical effect, see above, p.43. 306 On the territorial limits of a civitas, see Roux (1994). 307 That Pliny draws a distinction between settled and pastoralist Arabs is clear from NH 6.143-44 where the Skenitai are specifically noted as wandering (vagi) and the Nabataeans inhabit a town (“Nabataei oppidum incolunt Petram nomine...”), Ross (2001) 23. 94/448 understood as a collection of small villages with some sort of unitary political structure, analogous to the other ethnic units described as such in Northern Syria. 308 The Euphrates The third part of the Mesopotamian borderland which Pliny describes consists of a series of cities and peoples downstream from Commagene: ...oppida adluuntur Epiphania et Antiochia, quae Ad Euphraten vocantur, item Zeugma LXXII [m.]p. a Samosatis, transitu Euphratis nobile. ex adverso Apameam Seleucus, idem utriusque conditor, ponte iunxerat. 309 The latter two of these are well known, Zeugma (Seleucia) on the west bank of the Euphrates and Apamea on the east. 310 A few sections before, Pliny had referred to Zeugma as Seleucia-on-the-Euphrates (NH 5.82) without noting their correspondence. 311 Kennedy argues that this is evidence of interchangeable use of the two names, and while that may be the case, it is not clear that Pliny realised that he was doing so. More likely he was simply following the names provided by his sources. The first two cities mentioned by Pliny are more difficult. The locations of Antioch ad Euphraten and Epiphaneia ad Euphraten are unknown. 312 Neither city is mentioned by any other literary sources, although coins of the second century CE survive from Antioch ad Euphraten. Grainger suggests that the two cities were another pair of twin cities on either side of a Euphrates crossing founded as such by Antiochus IV Epiphanes. 313 The site of Europus on the Euphrates has not been securely identified. Carchemish is the most likely location, but its location on the 308 Ando examines Strabo’s treatment of the social, cultural and legal status of a civitas in Ando (2012). 309 Pliny NH 5.86, quoted above, p.69. 310 Appendix 1.20. Cohen (2006) 190–96. 311 Cohen reports that the identification of the town of Zeugma with Seleucia ad Euphraten (NH 5.82) was confirmed by a dedicatory inscription of the late second or early third century by a decurion from Seleuciae Zeugmae, Cohen (2006) 190. The inscription, a dedication to Jupiter Dolichenus, a cult derived from an ancient sanctuary at Doliche in Commagene, was found in Pannonia at Brigetio (modern Ószöny), the location of one of several European cult sites (called Dolichena): Hörig and Schwertheim (1987) 155–6 no. 236; finds: nos. 237– 261 (perhaps no. 262) 312 Antioch on the Euphrates: Cohen (2006) 151–52. Epiphaneia on the Euphrates: Cohen (2006) 169. 313 Grainger (1990) 138. 95/448 modern border between Syria and Turkey has hindered archaeological exploration of the site. 314 Amongst these cities, Pliny notes two other groups. The first is a people called “Rhoali”, who dwell next to Mesopotamia. 315 They are not mentioned elsewhere and Pliny gives no further information. 316 Retsö characterises Dillemann’s suggestion that Rhoali could be a distortion of Osrhoeni as a guess. 317 The second group mentioned by Pliny are the Arab Skentitai who are mentioned in the grammatically confused passage discussed above: “at in Syria oppida Europum, Thapsacum quondam, nunc Amphipolis; Arabes Scenitae.” 318 From the known sites in Pliny’s narrative, the narrative logic seems to be a progression of sites following the river downstream. If so, Pliny places the lands of the Skenitai Arabs downstream of the Europus/Thapsacus/Amphipolis group (however that was constituted), and upstream of Sura. Ecologically , this is a likely area for pastoralism. As the Euphrates flows south away from the hills, rain becomes less frequent and thus the land more marginal and less suited to dryland farming. 319 However, in the passage describing the course of the Euphrates through the Taurus, Pliny demonstrated a willingness to be flexible with the relationship between his narrative and the geographic reality of the spaces he describes. Thus, caution is required, especially as many of the other sites in this section of Pliny’s work have yet to be securely located. 314 See Appendix 1.11. Charchemish is one of the most likely sites for Thapsacus, see Appendix 1.18. 315 Pliny NH 5.87: “qui cohaerent Mesopotamiae, Rhoali vocantur.” 316 W eissbach’s entry in Pauly-Wissowa (‘Rhoali’ RE) notes the presence of a north Arabian tribe with the name “Ruwala” (Lancaster, W . and Felicity , s.v . EI). 317 Dillemann (1962) 77; Retso (2003) 428, n.177. 318 Pliny NH 5.87. The section is quoted in full at p.58. 319 Xenophon calls this area Arabia (“διὰ τῆς Ἀραβίας,” Anab. 1.5.1). The Arab Skenitai and nomadic pastoralism will be discussed in Chapter 5. 96/448 Further down the Euphrates, Pliny’s narrative arrives at Sura just after the river bends to the east. 320 Pliny is the first to mention this site, which seems to have become more important in late antiquity when it became the site of a legionary base near the northern end of the Strata Diocletiana. 321 After digressing from the Euphrates to discuss the desert cities, Pliny’s narrative returns to the river at Sura and quickly proceeds 320 Pliny NH 5.87: “ita fertur usque Suram locum, in quo conversus ad orientem relinquit Syriae Palmyrenas solitudines, quae usque ad Petram urbem et regionem Arabiae Felicis appellatae pertinent.” Translated above, p.69. Kessler, Karlheinz ‘Sura [3]’ BNP . This is a different Sura to the major centre of Babylonian Judaism in Southern Babylonia (Ego, Beate, ‘Sura [4]’ BNP . 321 Kessler, Karlheinz, ‘Sura’ BNP . For more on the military road usually ascribed to Diocletian, see Whittaker (1994) 135–39; Isaac (1990) 163–71; several of the forts of the Strata Diocletiana are discussed in Kennedy (1990). 97/448 Map 12: Pliny's Euphrates Itinerary downstream to Babylonia. In the same way as Strabo, Pliny chooses not to describe the various settlements and stopping points further down the Euphrates that we see in Isidore’s itinerary . 322 Pliny only mentions Philiscum, an “oppidum Parthorum ad Euphraten”. This is probably the same place which Isidore calls Phaliga (Φάλιγα). 323 Isidore’s narrative treats this site with some importance, claiming that the name means “halfway station” (μεσοπορικόν) in Greek, giving distances measurements to Zeugma and Seleucia on the Tigris. Accordingly , it probably corresponds with the Aramaic name Pelga (Syriac ܐܓܠܦ meaning half or middle) in the vicinity of Circesium. 324 Palmyra The final part of Pliny’s sequence centres on Palmyra. This was an important borderland city for which Pliny’s testimony is among the few extant literary descriptions. This passage recurs frequently in my analysis of the Mesopotamian borderland. In this section, I discuss its geographical implications. At Sura, Pliny digresses from the Euphrates to describe Palmyra and the other cities facing the Syrian desert. These are part of the structure that Jones imagined when he described Syria as two lines of port cities: one on the Mediterranean coast, the other on the edge of the desert. 325 After describing Palmyra itself, Pliny draws the line of desert ports verbally: Infra Palmyrae solitudines Telendena regio dictaeque iam Hierapolis ac Beroae et Chalcis. ultra 322 Isidore 1; Strabo 16.1.27. 323 Isodore 1: “εἶτα Φάλιγα κώμη πρὸς τῶι Εὐφράτηι (λεγοιτο δ᾽ ἂν ῾Ελληνιστι μεσοπορικόν), σχοῖνοι ς\. ἀπὸ ᾽Αντιοχειας ἕως τούτου σχοῖνοι ρκ· εντεῦθεν δὲ επι Σελεύκειαν την πρὸς τῶι Τιγριδι σχοῖνοι ρ\. παράκειται δὲ τῆι Φάλιγα κωμόπολις Ναβαγάθ, καὶ παραρρεῖ αὐτὴν ποταμὸς ᾽Αβούρας, ὃς ἐμβάλλει εἰς τὸν Εὐφράτην·” (Then Phaliga, a village near the Euphrates (it is said to mean “halfway station” in Greek), 6 schoinoi. From Antioch to here is 120 schoinoi, and then to Seleucia on the Tigris is 100 schoinoi. Lying near Phaliga is the small town of Nabagath and flowing nearby is the Abouras river, which goes into the Euphrates.) 324 Pliny 5.89: “A Sura autem proxime est Philiscum, oppidum Parthorum ad Euphraten. ab eo Seleuciam dierum decem navigatio, totidemque fere Babylonem.” (But next from Sura is Philiscum, a Parthian town on the Euphrates. From there it is ten days voyage to Seleucia and just about as many to Babylon.) For Pelga and Circesium, see Appendix 1.8. 325 Jones (1971) 227. This metaphor is common in the literature on the so-called “caravan cities” of the region, Millar (1998) 119. 98/448 Palmyram quoque ex solitudinibus his aliquid obtinet Hemesa, item Elatium, dimidio propior Petrae quam Damascus. 326 This description places Palmyra within the context of other important cities on the eastern fringe of Roman control, from Hierapolis in Roman Syria near the border of Commagene, to Petra, capital of the Nabataean kingdom near the far southern extension of the rift valley (Map 13: Desert Ports in Pliny). The regio Telendena is otherwise unknown and cities to the north (Hierapolis, Beroae and Chalcis) have already been discussed. Elatium is also unknown, while Emesa, Damascus and Petra are well known, but beyond my study area. Palmyra is the borderland city which receives the most description in the Natural History. Of particular note is the position Pliny assigns the city between Rome and Parthia. Palmyra urbs nobilis situ, divitiis soli et aquis amoenis, vasto undique ambitu harenis includit agros, ac velut terris exempta a rerum natura, privata sorte inter duo imperia summa Romanorum Parthorumque, et prima in discordia semper utrimque cura. Abest ab Seleucia Parthorum, quae vocatur Ad Tigrim, CCCXXXVII [m.]p., a proximo vero Syriae litore CCIII et a Damasco XXVII propius. 327 Pliny places Palmyra in three ways: physically amidst a “vast circuit of sand”, politically “between two empires” and mathematically in terms of distances to a number of important points. Several aspects of this passage will be discussed fully in subsequent chapters, 328 but it should be noted now that all three frameworks reflect the actual situation imperfectly and illustrate Pliny’s representation of the space. I will discuss the physical aspects of Pliny’s description here. 326 Pliny NH 5.89: “On this side of the desert of Palmyra is the Telendena region, and Hierapolis, Beroea and Chalcis, already mentioned. On the other side of Palmyra, Emesa holds some of this desert, as does Elatium, which is half as close to Petra as Damascus.” 327 Pliny , NH 5.88: “Palmyra is a city famous for its position, the richness of its soil, and the quality of its water, its fields surrounded on all sides by a vast circuit of sand, as if cut off from the world by nature itself, a private lot between two great empires of Rome and Parthia, and at the first sign of discord between them, always a concern to both. It is distant 337 miles from Seleucia of the Parthians, generally known as Seleucia on the Tigris, 203 from the nearest part of the Syrian coast, and twenty-seven less from Damascus.” 328 Palmyra’s murky political and legal relationships to Rome and Parthia will be addressed in Chapter 2. The city’s important role in networks of trade and contact will be examined in Chapter 5. 99/448 Pliny notes the specific local advantages for which Palmyra was renowned: its position and the quality of its soil and water. The latter two were no doubt more remarkable to a Roman audience because of the first. Palmyra was situated at an oasis in a valley roughly halfway between Damascus and the Euphrates, from which it was well placed to exploit several desert trade routes between Syria and Babylonia. 329 Pliny alludes to that web of trade routes surrounding Palmyra when he gives the location of Palmyra in terms of distances to other important spaces of contact: Seleucia on the Tigris, Damascus, and the Mediterranean (in 329 The area around Palmyra was considerably more productive and less isolated than Pliny implies (Gawlikowski (1983) 58–59), although the city’s rapid growth probably required Palmyra to import food, perhaps from Dura Europus (Dirven (1996) 41–42, 45–46). Smith (2013) provides a detailed examination of Palmyrene history and society. Hekster and Kaizer (2004) note that Pliny’s description is suspiciously similar to what an oasis should look like. 100/448 Map 13: Desert Ports in Pliny the form of the Syrian coast, Syriae litore). As one might expect, there is some inaccuracy in these figures. The straight-line distance between Seleucia and Palmyra is actually 405 m.p., not 337 m.p., and the distance between Damascus and the coast is 57 m.p. rather than 27 m.p.. The distance from Palmyra to the nearest part of the Syrian coast is 141 m.p., not 203 m.p. as Pliny gives it, but if one travels first to Damascus and then to the nearest part of the coast, the distance is 202 m.p.. However, the correspondance of this adjusted measurement is coincidental; Pliny’s measurements reflect the reality of human and animal movement over natural terrain, measured in days of travel, then converted into Roman miles, not precise mathematical calculations based on coordinate geometry. 330 Book Six: Mesopotamia Pliny’s general procedure in book five was to describe Roman space according to the Roman system of organisation, that is, province by province. In his descriptions of the non-Roman spaces of Asia in book six his approach changes, not to describe space according to non-Roman systems of organisation, but to follow the features of physical geography . This reflects the ideological division between Roman and Parthian space to which I referred earlier and which will be fully explored in Chapter 4. As part of a periplus around the outside of Asia, his narrative returns to Mesopotamia. Oppida praeter iam dicta habet Seleuciam, Laodiceam, Artemitam; item in Arabum gente qui Orroei vocantur et Mandani Antiochiam quae a praefecto Mesopotamiae Nicanore condita Arabis 331 vocatur. [118] Iunguntur his Arabes introrsus Eldamari, supra quos ad Pallacontam flumen Bura oppidum, Salmani et Masei Arabes; Gurdiaeis vero iuncti Azoni, per quos Zerbis fluvius in Tigrim eadit, Azonis Silices montani et Orontes, quorum ad occidentem oppidum Gaugamela, item Suae in rupibus. Supra Silicas Sitrae, per quos Lycus ex Armenia fertur, ab Sitris ad hibnernum exortum Azochis oppidum, mox in campestribus oppida Dios Pege, Polytelia, Stratonicea, Anthemus. [119] In vicinia Euphratis Nicephorion, quod diximus; Alexander iussit condi propter loci opportunitatem. Dicta est et in Zeugmate Apamea; ex qua orientem petentes excipit oppidum Caphrena munitum, quondam stadiorum LXX amplitudine et Satraparum Regia appellatum quo tributa conferebantur, nunc in arcem redactum. [120] durant, ut fuere, Thebata et, ductu Pompei Magni terminus Romani 330 Although the mathematical tools for calculating distances on a coordinate plane were available, Berggren, Jones, et al. (2000) 31–41. 331 Arabis (or Arabes) is a conjecture for “Arabs” in the manuscript tradition. 101/448 imperi, Oruros, a Zeugmate CCL. 332 While Pliny generally organises his narrative of non-Roman Asia according to geographical features, his description of Mesopotamia itself in book six is vague about the specific topology . The cities and peoples in this passage are sparsely described and are usually located with reference to rivers and each other but not in a clearly two dimensional way . In fact, the passage is structured as a series comprising a jumbled mixture of lists of undescribed cities which are not located in space, individual cities with brief descriptions (sometimes located in space) and sets of sequentially located spaces and groups which give the impression of a traveller’s narrative. The spaces in the series are related to each other only by implication and context. Thus Pliny’s description of Mesopotamia in book six is structured as a list, within which further lists can be found. Pliny’s Mesopotamian Lists The first of these spatially linked lists describes the Arab tribes that inhabit Mesopotamia proper (Map 14: Pliny's Mesopotamian Tribes): item in Arabum gente qui Orroei vocantur et Mandani Antiochiam quae a praefecto Mesopotamiae Nicanore condita Arabis vocatur. [118] Iunguntur his Arabes introrsus Eldamari, supra quos ad Pallacontam flumen Bura oppidum, Salmani et Masei Arabes. 333 Of these tribes, only the Orroei were mentioned by Pliny in book five. 334 Mandani is likely to be an early 332 Pliny NH 6.117-120: “Except for the towns already mentioned, [Mesopotamia] contains Seleucia, Laodicea, Artemita; and in Arabia, the people called the Orroei, and Antiochia of the Mardani, founded by Nicanor, the governor of Mesopotamia, and called Arabis*. [118] Joined to these in the interior are the Eldamari Arabs (beyond whom is the town of Bura near the river Pallaconta) and the Salmani and Masei Arabs. Next to the Gordyaei are the Azoni, through whose territory the river Zerbis runs into the Tigris; next to the Azoni are the Silici, a mountain tribe, and the Orontes, to the west of whom lies the town of Gaugamela, as also Suae among the rocks. Beyond the Silici are the Sitrae through whose district the river Lycus flows out of Armenia, south- east of Sitrae the town of Azochis, then in the plains the towns of Dios Pege (Zeus’ Spring), Polytelia, Stratonice, and Anthemus. [119] In the vicinity of the Euphrates is Nikephorion, which we have mentioned; Alexander ordered it to be built because of the favourableness of the site. Apamea at Zeugma has been mentioned also, from which, heading east, one comes to Caphrena, a fortified town, formerly seventy stadia in size and called the "Palace of the Satraps." Tribute used to be brought here; now it is just a fortress. [120] Thebata is still in the same state as formerly , and Oruros, the limit of Roman power under Pompeius Magnus, 250 miles from Zeugma.” 333 Pliny NH 6.117-18, translated above, p.67. 334 Pliny NH 5.85, discussed above. 102/448 scribal error for Mygdones, where a cursive ΥΓ has become ΑΝ. 335 Thus the Mandani Antiochia founded by Nicanor is Strabo’s Mygdonian Antiocheia, that is, Nisibis. 336 The location of the Arab tribes listed by Pliny is difficult to ascertain. Pliny’s Pallaconta river is the Pollacopas canal, which ran south from Pirisabora (Anbar) to the Arabian Gulf. 337 The canal is thought to have given its name to the modern town al-Fallūǧa 335 Thanks to Kevin Van Bladel for the suggestion. 336 Strabo 16.1.23; see Appendix 1.14. Mygdonian Antiochia does appear in Pliny’s work by that name, see below , p.109ff. 337 Bivar (2000) 69–70; Retso (2003) 406; Meissner (1896). Alexander sailed south on the Pallaconta during his preparations for his proposed Arabian campaign, Arr. Anab. 7.21. For Pirisabora/Anbar: Oelsner, Joachim, ‘Ambarra’ BNP; Matthews (1989) 132, 149, 174–75; Musil (1927) 353–57. 103/448 Map 14: Pliny's Mesopotamian Tribes (Fallujah) which lies near the ruins of Anbar. Retsö suggests an identification for Bura from a town near Baghdad mentioned by the thirteenth-century geographer Yaqut al-Hamawi. 338 This would place the Eldamari Arabs along the north bank of the Euphrates between Osrhoene and Pallaconta/Fallūǧa and the Salmani and Masei further south and east. 339 The second such list begins with reference to the Gordyaei in the mountains to the north of Mesopotamia (see Map 8: The Skenitai and their Lands). Gordyaeis vero iuncti Azoni, per quos Zerbis fluvius in Tigrim eadit, Azonis Silices montani et Orontes, quorum ad occidentem oppidum Gaugamela, item Suae in rupibus. Supra Silicas Sitrae, per quos Lycus ex Armenia fertur, ab Sitris ad hibnernum exortum Azochis oppidum, mox in campestribus oppida Dios Pege, Polytelia, Stratonicea, Anthemus. [119] In vicinia Euphratis Nicephorion, quod diximus; Alexander iussit condi propter loci opportunitatem. 340 The key locations that anchor this section to the left bank of the Tigris are Gaugamela, near Arbela, the Lycus, one of the rivers bounding the geographical region of Adiabene, and the Gordyaei. 341 The “Gordyaei” are Strabo’s Gordyaeans and Xenophon’s Carduchians, who inhabit the mountains between Armenia and Mesopotamia, probably around the Upper Tigris Basin on both sides of the river until it descends into the plains of Mesopotamia near the modern city of Cizre. 342 The Azones inhabit the lands on either side of the “Zerbis” river. Olshausen assumes that “Zerbis” refers to either the Greater or Lesser Zab, the former also known in antiquity as the Lycus (modern al-Zāb al-Kabīr) and the latter as the Caprus (modern al-Zāb al- Ṣaġīr). 343 However, the peoples of this passage are well linked topologically , progressing from the Gordyaeans 338 Retso (2003) 406. Baghdad lies on the Tigris, due west of Fallujah. 339 Retso (2003) 406 speculates that Masei may be related to Mesene in some way: “The Salmani and Masei are difficult to locate, unless it is assumed that Masei has something to do with Mesene. It could thus refer to the kingdom of Charax and support the assumption of Arab presence in that area.” 340 Pliny NH 6.118-19, translated above, p.67. 341 The existence of a kingdom of Adiabene that expanded beyond this core area means that Adiabene sometimes has a wider, political, meaning in the sources. 342 Pliny NH 6.129. Pliny elsewhere calls them Carduchi and Cordueni (NH 6.44); the variation may mean simply that Pliny has derived his information from different sources. For Strabo’s Gordyaeans, see 68ff. 343 Olshausen, Eckart, ‘Zerbis’ BNP . 104/448 to the northwest, and moving south along the east bank of the Tigris through Adiabene. Pliny lists the Zerbis before the Lycus/Greater Zab, so unless Pliny’s account is garbled, the Zerbis should lie to the north of the Lycus, whereas, the Caprus/Lesser Zab lies south of the Lycus. For this reason, Marciak suggests that Pliny’s Zerbis be identified with the Botan; 344 but that river (also known by Classical authors as the Kentrites) which separated the Gordyaeans from the Armenians, should lie to the north of the Gordyaeans, not to the south. The Zerbis could be the Little Khabur, a tributary of the Tigris between the Botan and the Greater Zab that enters the Tigris about 25 kilometres south of Cizre. This would place the Azones in roughly the same area as I have suggested for Strabo’s Chazene (Χαζήνη). 345 Next to the Azones are the Silices and the 344 Marciak (2011) 186–87. 345 See Appendix 1.1. 105/448 Map 15: Pliny's Mountain Tribes Orontes: the Silices are explicitly a mountain tribe, while the Orontes are east of Gaugamela. The grammar is unclear regarding Suae, which could either be another tribe, or a town or place used as a topological reference point. The Sitrae lie around the river Lycus (the Greater Zab), further south or east from the Silices (supra) and northeast of a town called Azochis. Pliny does not associate these groups directly with the towns that follow , Dios Pege (Zeus’ Spring), Polytelia, Stratonicea, and Anthemus, rather he connects those towns and the topologically related groups with an ambiguous “mox” (next). The towns are simply located “in the plains” (in campestribus oppida). Which plains? Except for Anthemus, none of these names are known in Mesopotamia or Adiabene, and Athemusia lies between Edessa and the Euphrates. While the known city of Anthemus is far to the west of the preceding peoples, the very next sentence picks up in eastern Mesopotamia at Nikephorion. Pliny previously placed Nikephorion in his praefectura Mesopotamiae, but without comment. 346 In this second reference to the site, he adds that it was founded by Alexander because of the advantages of the site. 347 In both books, Pliny mentions Anthemus/Anthemusia in close proximty with Nikephorion. Is this the same Anthemus, or a different town of the same name located in Adiabene where the preceding narrative leads? The link to Nikephorion suggests that they are both the same, well-known, site between Zeugma and the Bailkh river. However, in the absence of any direct evidence to anchor the other three oppida to any specific part of the region through which Pliny’s narrative meanders, these towns cannot be located. After the second list, Pliny returns to the banks of the Euphrates to give longer descriptions of some western Mesopotamian locations (Map 11: Mesopotamian Sites in Pliny NH 5 & Map 16: Mesopotamian Sites in Pliny NH 6). The structure of this final list gives every impression of an itinerary fragment. Dicta est et in Zeugmate Apamea; ex qua orientem petentes excipit oppidum Caphrena munitum, quondam stadiorum LXX amplitudine et Satraparum Regia appellatum quo tributa conferebantur, nunc in arcem redactum. [120] durant, ut fuere, Thebata et, ductu Pompei Magni terminus Romani 346 Pliny NH 5.86, quoted and translated at p.52. 347 Alexander is one of several contenders for the honour of this foundation; see Appendix 1.13. 106/448 imperi, Oruros, a Zeugmate CCL. 348 Pliny mentioned Apamea in conjunction with his narrative of the Euphrates in book five. 349 Here Apamea begins a new itinerary within Pliny’s narrative, as indicated by the beginning of the next sentence: “ex qua orientem petentes excipit” (“leaving that city towards the east...”). The destination is the fortified town (munitum) of Caphrena, which Pliny reports to once have been a large centre for tribute collection known as Satraparum Regia. The name could be a reference to Persian or Macedonian administration and such a site would attest to the importance of Mesopotamia within those empires. 350 However, the size of the space (70 stadia!), suggests that Caphrena was associated with a Persian hunting park, a παράδεισος or paradisus, rather than an enormous and now vanished city . 351 Several candidates have been proposed for the site of Caphrena, but none are secure. 352 After Caphrena, Pliny mentions Thebata with the curious note that it “remains unchanged” (“durant ut fuere”). Pliny had not previously mentioned a town named Thebata. An earlier reference to Thebata may have been lost from the Natural History itself, perhaps from book five. Pliny may have copied this phrase from his source, without also copying the antecedent. Or he may be drawing a contrast between the multiple 348 Pliny NH 6.119-120. Translated above, p.67. 349 Pliny NH 5.86. 350 Recalling Herodotus’ taxation list (3.92) which placed the tax value of “Babylonia and the rest of Assyria” at 1000 talents of silver and 500 castrated boys. Quoted above, p.42. 351 Dillemann (1962) 170; Briant (2002) 201–2, 442–46; Ego, Beate ‘Paradise’ BNP . As Wilkinson (2005) 50 points out such hunting parks probably left a light impression on the physical landscape and as such resist location and identification. 352 The editors of Pleiades confidently locate Caphrena at Halfeti on the banks of the Euphrates four kilometres downstream of Rumkale: ‘Caphrena’ Pleiades (http://pleiades.stoa.org/places/658424, accessed 21 March 2013). Dillemann (1962) 170. The Barrington Atlas offers an alternative, namely Ank Köy , a site some five kilometres east of the river and seven from Halfeti: BAtlas 67 F2 (For Caphrena the directory notes: “Ank Köy?”). Dillemann ((1962) 170) proposes that Satraparum Regia, the alternate name given for Caphrena by Pliny , is a translation of the Greek Basileia and Persian Apadna. Apadna is found as a toponym elsewhere in Mesopotamia, at the modern site of Tell Harzem in the Ghars valley , around 10km (70 stades) from the source of the Ghars river at Ras el ‘Aïn (Dillemann (1962) 159). Finally , Sinclair’s catalogue of east Turkish sites refers to the ruins of a town called Kafrhan located north or northwest of Edessa, last mentioned in a traveller’s account of 1911 (Sinclair (1987) 4.179). 107/448 names and lives of Caphrena/Satraparum Regia and this more durable town. At any rate, Thebata is otherwise unattested, although Dillemann suggests that it may be the same place that Ammianus calls Thilsaphata. 353 Finally , Pliny mentions another site, Oruros, which he calls the boundary of the the empire (“terminus Romani imperi”) in Pompey’s time. Pliny places this 250 Roman miles (370 km) from Zeugma. If measured in a straight line, this would put Oruros, and thus Pompey’s border, in the vicinity of an arc linking Dura Europus, Singara and Pinaka where the Tigris emerges from the mountains. In other words, 353 Dillemann (1961) 311–12. Amm. Marc. 25.8.16. 108/448 Map 16: Mesopotamian Sites in Pliny NH 6 just beyond the Khabur river and the Tur Abdin range. This matches well with the later historical accounts of the first-century CE campaigns of Lucullus and Pompey . 354 It also matches well with the general location of the fortress Ammianus calls Ur. 355 Dillemann suggests that Ur/Oruros be located at modern ‘Aïn el Chahid (Arabic ‘aïn, Syriac ܐܢܝܥ, spring) where Sarre and Herzfeld saw the remains of a Roman camp. 356 This itinerary marks the end of Pliny’s description of northern Mesopotamia. Pliny next jumps to a canal near Babylon and proceeds to describe the southern region. As in book five, the structural logic of this part of book 6 is the list. In Chapter 4, I will argue that the chaotic organisation of these lists (one of Arab tribes, one of mountain tribes, one of dislocated cities, and one short itinerary fragment) contrasts with Pliny’s orderly presentation of Syria in book 5. Adiabene and Parapotamia Pliny makes several references to Adiabene throughout his work. 357 In his broad initial description of the various lands that constitute “Syria”, he includes Adiabene, “Assyria ante dicta” (once called Assyria). 358 W e have seen that the geographical bounds of “Assyria” are flexible; here Pliny gives only the topological note that it lies beyond Armenia (“ultra Armeniam”). 359 Two passages give further, more detailed, topological information, and illustrate the flexibility of the terms “Adiabene” and “Assyria”. The first comes in Pliny’s description of the borders of Armenia where Pliny notes that the Tigris and impassable mountains enclose Adiabene, “Adiabenen Tigris et montes invii cingunt.” 360 This is an accurate description of the geographical region, framed as it is by the Tigris river and the Zagros mountains. Pliny might also have mentioned the 354 Lucullus’s siege of Nisibis: Dio 36.6-7. Pompey’s intervention in a dynastic squabble in Gordyene: Dio 37.5.2-5. Pompey’s forces marched through Mesopotamia, aided by local communities, even if the route was “contrary to the agreement with the Parthian” (παρὰ τὰ συγκείμενα πρὸς τὸν Πάρθον). 355 Amm. Marc. 25.8.6-7. 356 Dillemann (1961) 311; Sarre (1911) 2.305–307. 357 Pliny HN 5.66; 6.25; 6.28; 6.41-42; 16.44. 358 Pliny NH 5.66, see above, p.49. 359 See above, p.53ff. Adiabene borders Armenia in the north at Pliny NH 6.25. 360 Pliny NH 6.28. 109/448 Greater and Lesser Zab rivers, but that would be beyond the general level of precision required for a passage that aims to describe the bounds of Armenia, rather than Adiabene itself. The second passage is more detailed, and more problematic. Namque Persarum regna, quae nunc Parthorum intellegimus, inter duo maria Persicum et Hyrcanium Caucasi iugis attolluntur. utrimque per devexa laterum Armeniae Maiori a frontis parte, quae vergit in Commagenen, Cephenia, ut diximus, copulatur eique Adiabene, Assyriorum initium, cuius pars est Arbilitis, ubi Darium Alexander debellavit, proxime Syriae. [42] totam eam Macedones Mygdoniam cognominaverunt a similitudine. oppida Alexandria, item Antiochia quam Nesebin vocant; abest ab Artaxatis DCCL [m.]p. fuit et Ninos, inposita Tigri, ad solis occasum spectans, quondam clarissima. 361 Pliny places Adiabene adjacent (proxime) to Syria, a statement which requires that Mesopotamia be aggregated to one of the two regions. Pliny’s list of cities in Adiabene shows part of this; Adiabene contains Alexandria, Antioch (“quam Nesebin vocant”), and Ninos. The description of Nineveh, under the well- known alternate name of Ninos, fits perfectly . Alexandria of Mygdonia is thought to lie near Arbela. It was probably founded in 331 BCE to commemorate the Macedonian’s famous victory at nearby Gaugamela, as Pliny’s note on Arbelitis recalls (“Arbilitis, ubi Darium Alexander debellavit”). 362 While Ninos and Alexandria of Mygdonia both lie within the area on the east bank of the Tigris that we might expect Adiabene to encompass, the last does not. Antioch quam Nesebin vocant, that is, Nisibis (Mygdonian Antioch) lies to the west of the Tigris in a region not otherwise attested as Adiabene. In fact, Pliny seems to equate Adiabene and Mygdonia (“totam eam Macedones Mygdoniam cognominaverunt...”), although he specifies the region he refers to as Mygdonia no more closely than does Strabo. In his analysis of Seleucid- Parthian Adiabene in geographical and ethnographical texts, Michal Marciak resolves the inclusion of 361 Pliny NH 6.41-42: “The kingdom of the Persians, which we now know as that of the Parthians, was established on the heights of the Caucasus mountains, between two seas, the Persian and Hyrcanian. As I said, Cephenia is joined to Greater Armenia on both sides by steep slopes running towards the front part which faces Commagene, and to this, Adiabene, where the land of the Assyrians begins, of which the part closest to Syria is Arbelitis, where Alexander defeated Darius. The Macedonians have named this all Mygdonia from its similarity [to Mygdonia in Macedon]. Its towns are Alexandria and Antiochia, which they call Nesebis; it is 750 m.p. from Artaxata. Ninos (Nineveh), placed on the Tigris looking to the west was once very famous.” 362 Dillemann (1962) 160, n.3; ‘Alexandrian Foundations (1)’ PECS, p.39; Cassius Dio mentions the location in his description of Trajan’s invasion: Dio 68.26; Theophylactus (5.7.10-11) calls it Alexandriana. 110/448 Nisibis within Adiabene by suggesting that Pliny is here writing about the kingdom of Adiabene, a political rather than geographical entity . 363 Marciak notes two important geographical notes provided by Josephus: the gift of Gordyene from the king of Adiabene to his son between 22 and 30 CE and the gift of Nisibis from the Parthian king Artabanos II to that same son, now king of Adiabene. 364 Thus in Pliny , the kingdom of Adiabene extends as far as Nisibis, the lands of which are adjacent (proxime) to Syria. In this passage, Pliny is likely to be using “Syria” synonymously with Assyria, or rather terrae Assyriorum, the lands of the Aramaic- speaking Assyrians, as he does in his initial survey of the space at NH 5.66, rather than in the political sense of the Roman provincia. 365 The final area to consider before leaving Pliny’s description of the borderland is Parapotamia. In book six, Pliny refers to a region by this name, but it is not the same area as that described by Strabo. 366 Proxima Tigri regio Parapotamia appellatur. In ea dictum est de Mesene; oppidum eius Dabitha; iungitur Chalonitis cum Ctesiphonte, non palmetis modo verum et olea pomisque arbusta. Ad eam pervenit Zagrus mons ex Armenia inter Medos Adiabenosque veniens supra Paraetacenen et Persida. Chalonitis abest a Perside CCCLXXX [m.]p.; tantum a Caspio mari et a Syria abesse conmpendio intineris aliqui tradunt. [132] inter has gentes atque Mesenen Sittacene est, eadem Arbelitis et Palaestine dicta. oppidum eius Sittace Graecorum, ab ortu et Sabdata, ab occasu autem Antiochia inter duo flumina Tigrim et Tornadotum, item Apamea, cui nomen Antiochus matris suae inposuit; Tigri circumfunditur haec, dividitur Archoo. [133] Infra est Susiane... 367 Pliny uses “Parapotamia” to refer to the land near the Tigris, rather than near the Euphrates, as Strabo does. Pliny explicitly includes Mesene within this area, and maybe also Chalonitis, Sittacene and Arbelitis (the 363 Marciak (2011) 191–94. 364 Marciak (2011) 192. Gordyene: Josephus AJ 20.24. Nisibis: Josephus AJ 20.68. 365 See above, p.80ff. with Map 9: Pliny's Syrian Overview . 366 See above, p.41. 367 Pliny NH 6.131-133: “The region next to the Tigris is called Parapotamia. Mesene, mentioned above, is in this area; it has a town called Dabitha. Adjoining Mesene is Chalonitis, with Ctesiphon, an area forested not only with palms, but also olives and fruit trees. The Zagros mountains reach this place, coming out of Armenia between Media and Adiabene, above Paratacene and Persis. Chalonitis is 380 miles from Persis; some say it is the same distance from the Caspian sea and from Syria by the shortest routes. [132] Between these people and Mesene is Sittacene, also called Arbelitis and Palaestina. Sittace is a Greek town of this region, to the east is Sabdata and to the west is Antiochia between the two rivers, Tigris and Tornadotus, and Apamea, to which Antiochus gave the name of his mother; this [town?] is surrounded by the Tigris and divided by the Archous. [133] Below is Susiana...” 111/448 latter two of which Pliny conflates). Pliny does not mention the eponymous city of Chala (Halwan), only the important Parthian capital Ctesiphon. 368 Pliny understands Chalonitis to lie south of the Silla river (the modern Diyala) and Sittacene north of the same, adjacent to Arbelitis. Pliny’s Parapotamia thus lies generally south of Adiabene and does not correspond to the area by that name mentioned by Strabo. Conclusion Pliny’s narrative of the Mesopotamian borderland uses geographical information as a general framework for the region as a whole, not as an aid to a deeper knowledge of the topological relationships which exist within the space. His description of Syria shows that he prefers to use linear features like rivers and coastlines to establish those topological relationships. In his description of northern Mesopotamia, Pliny creates his own linear features by using lists and itineraries to relate places to one another. Compare this approach to Strabo’s practice of dividing space into successively smaller subdivisions, defined according to different criteria, as fits the subdivisions he defines. As can be seen in the case of Parapotamia, while Pliny and Strabo agree on most broad geographical points, they do not define the spaces of the Mesopotamian borderland in the same way . Neither author deliberately distorts the geographical reality of the region, but by defining the borders of the space as a whole and the subsidiary spaces that comprise the borderland according to their own conception of the region, they present a geographical image that shapes their political narrative of Romano-Iranian relations. The political border between the empires was particularly open for this kind of representation; neither author is precise in this regard. 369 Claudius Ptolemy Ptolemy’s Geography divides the world into sections for the purpose of constructing a map. He 368 He also connects the two at Pliny NH 6.122. 369 For a discussion of these representations, see Chapter 4. 112/448 structures the work such that a reader attempting to do so could begin with northwest in the upper-left corner of a map oriented with north at the top and draw the sections in the order he presents with as little as possible danger of smudging the previously drawn sections with the drawing hand. 370 The bulk of the data Ptolemy’s work presents comprises coordinate points to locate the many point, line and area features he includes. He makes no attempt to show the paths of linear terrain features or the edges of area features at any but the coarsest level; points are given for rivers at the source, the mouth, and major junctions (ἡ συναφὴ); the vertices provided for regions present them as basic geometric shapes; mountain ranges are sometimes described as linear features, sometimes simply by their mid-point (τὸ μέσον). Ptolemy is particularly careful to show topological relationships of adjacency between the region he is currently describing and the others in his work. In fact, he criticises his main source, Marinus of Tyre, for a number of relatively trivial omissions of such relationships. 371 Ptolemy reaches the fertile crescent in book five in which he describes the boundaries, features and settlements of the borderland areas, Syria Coele (5.15), Mesopotamia (5.18), Arabia Deserta (5.19), and Assyria (6.1), as well as the adjacent regions of Greater Armenia (5.13) and Babylonia (5.20). 370 Ptol. Geog. 2.1.4: “Προειλόμεθα δὲ τάξιν τοῦ περὶ τὴν καταγραφὴν εὐχρήστου πανταχῆ ποιούμενοι πρόνοιαν, τουτέστι καθ’ ἢν ἐπὶ δεξιὰ ποιησόμεθα τὰς μεταβάσεις, ἀπὸ τῶν ἤδη κατατεταγμένων ἐπὶ τὰ μηδέπω τῆς χειρὸς ἐκλαμβανομένης. Τοῦτο δὲ γένοιτ’ ἂν, εἰ γράφοιτο τά τε βορειότερα πρότερα τῶν νοτιωτέρων καὶ τὰ δυσμικώτερα τῶν ἀπηλιωτικωτέρων, ὅτι πρὸς τὰς τῶν ἐγγραφόντων ἢ ἐντυγχανόντων ὄψεις ἄνω μὲν ἡμῖν ὑπόκειται τὰ βορειότερα, δεξιὰ δὲ τὰ ἀπηλιωτικώτερα τῆς οἰκουμένης ἐπί τε τῆς σφαίρας καὶ τοῦ πίνακος.” (W e have chosen an order [of presentation] with forethought to convenience in the drawing of the map in every respect, namely progressing toward the right, with the hand proceeding from the things which had already been inscribed to those that have not yet [been inscribed]: this would be achieved by having the more northern [places] drawn before the more southerly ones, and the more western before the more eastern, because our convention is that “up” with respect to the map-makers’ or spectators’ view means “north,” and “right” means “east” in the oikumene, both on a globe and on a planar map.) Berggren and Jones (2000) trans. 371 Ptol. Geog. 1.16. 113/448 Mesopotamia The boundaries of Ptolemy’s Μεσοποταμίας θέσις (“Arrangement of Mesopotamia”) conform largely with that of previous Roman geographers. Ἡ Μεσοποταμία περιορίζεται ἀπὸ μὲν ἄρκτων τῷ ἐκτεθειμένῳ μέρει τῆς Μεγάλης Ἀρμενίας, ἀπὸ δὲ δύσεως τῷ ἐκτεθειμένῳ παρὰ τὴν Συρίαν τοῦ Εὐφράτου ποταμοῦ μέρει, ἀπὸ δὲ ἀνατολῶν τῷ παρὰ τὴν Ἀσσυρίαν μέρει τοῦ Τίγριδος ποταμοῦ τῷ ἀπὸ τοῦ πρὸς τῇ Ἀρμενίᾳ τμήματος μέχρι τῶν τοῦ Ἡρακλέους Βωμῶν... ἀπὸ δὲ μεσημβρίας τῷ λοιπῷ μέρει τοῦ Εὐφράτου παρὰ μὲν τὴν Ἔρημον Ἀραβίαν μέχρι πέρατος... παρὰ δὲ τὴν Βαβυλωνίαν μέχρι τῆς πρὸς τὸν Τίγριν συναφῆς καὶ τοὺς εἰρημένους Βωμοὺς... 372 Ptolemy gives Mesopotamia five edges: Greater Armenia, the Euphrates near Syria (παρὰ τὴν Συρίαν) to the west, the Euphrates to the south near Arabia Deserta (παρὰ τὴν Ἔρημον Ἀραβίαν), Babylonia, and the Tigris near Assyria (παρὰ τὴν Ἀσσυρίαν). Thus Ptolemy bounds Mesopotamia by the two rivers, defines the northern boundary as Armenia, and maintains a distinction between Upper Mesopotamia (Mesopotamia) and Lower Mesopotamia (Babylonia) as does Strabo (see Map 2: Strabo's Assyria). 373 Despite his stated aim of producing a macro-level geography of the entire world rather than a chorography of a narrow region comprising “even the most minute features”, 374 Ptolemy locates far more features within Mesopotamia than any other extant geographical source. However, his is a mathematical geography and he provides almost no description of the features that fill the spaces he describes. Within Mesopotamia, Ptolemy places two mountain features, τό Μάσιον (Mount Masius) and ὁ Σιγγάρας (the Jebel Sinjar) (5.18.2); two rivers following from their sources to the Euphrates, ὁ καλούμενος Χαβώρας (the Khabur) and ὁ καλούμενος Σαοκόρας ποταμὸς (5.18.3); six districts (τὰ χωρία), Anthemusia, Chalkitis, Gauzanitis, Axabene, [T]ingene [Singene?], Ankobaritis (5.18.4); and sixty-nine cities and villages (πόλεις 372 Ptol. Geog. 15.18.1: “Mesopotamia is bounded on the north by the region of Greater Armenia already mentioned, on the west by the part of the Euphrates river near Syria (previously described), on the east by the part of the Tigris river near Assyria, from there to the part of Armenia as far as the altars of Heracles... To the south, the remaining part of the Euphrates near the Arabian Desert as far as navigable... And from Babylon up to the connection to the Tigris and the aforementioned altars...” Except where they are directly relevant, my quotations from Ptolemy will omit his coordinates. 373 Strabo 16.1.21. Pliny does not define the boundary between Mesopotamia and Babylonia. 374 Ptol. Geog. 1.1. See Introduction. 114/448 καὶ κῶμαι) (5.18.6-13). These settlements are divided into four groups: those near the Euphrates (5.18.6-7), those near the Tigris (5.18.9), Seleucia (5.18.8) and two neighbouring places (5.18.9), and those in the middle (5.18.10-13). Each of these geographical features is accompanied by at least one coordinate obtained in one of two ways: by astronomical observation by scientists, or by distance measurements obtained from travellers. 375 Ptolemy claims to grant priority to the astronomically derived coordinate data, but he admits that they are fewer and he does not distinguish between the two sources in the coordinate lists themselves. 376 Moreover, there is no guarantee that the “scientifically observed” data would be any more accurate than that calculated from travel times. In his introduction, Ptolemy spends several paragraphs describing his method for correcting the measurements deriving from travellers’ reports, but those methods ultimately come down to guesswork on Ptolemy’s part. 377 Nevertheless Ptolemy takes every pain to make his method clear. In his introduction, he gives frank consideration to the potential for error inherent in his coordinate derivation methodology . 378 Another source of error is that of scribal transcription. Numerals are notoriously prone to transcription errors under the best conditions, but the simple fact that much of Ptolemy’s text comprises 375 Ptol. Geog. 1.2. 376 Ptol. Geog. 1.4. 377 Ptol. Geog. 1.8-9. Rihll (1999) 100–101: “His finished work gives a spurious air of accuracy , however, because he decided to present all the material to hand – and not just the carefully observed data – in a digital fashion. For example, when told that the journey from X to Y took ‘10 or 12 days’ to cover, Ptolemy translated this into degrees of latitude or longitude via a thumb-rule of average number of stadia covered per day , and made other adjustments which in general seem to arise primarily from a desire to simplify the computations.” 378 Ptol. Geog. 2.1.2: “...ὅτι τὰς μὲν τῶν τετριμμένων τόπων μοιρογραφίας μήκους τε καὶ πλάτους ἐγγυτάτω τῆς ἀληθείας ἔχειν νομιστέον διὰ τὸ συνεχὲς καὶ ὡς ἐπίπαν ὁμολογούμενον· τῶν παραδόσεων· τὰς δὲ τῶν Ἀρξόμεθα δ’ ἐντεῦθεν τῆς κατὰ μέρος ὑφηγήσεως ἐκεῖνο προλαβόντες, ὅτι τὰς μὲν τῶν τετριμμένων τόπων μοιρογραφίας μήκους τε καὶ πλάτους ἐγγυτάτω τῆς ἀληθείας ἔχειν νομιστέον διὰ τὸ συνεχὲς καὶ ὡς ἐπίπαν ὁμολογούμενον· τῶν παραδόσεων· τὰς δὲ τῶν μὴ τοῦτον τὸν τρόπον ἐφοδευθέντων, ἕνεκεν τοῦ σπανίου καὶ ἀδιαβεβαιώτου τῆς ἱστορίας ὁλοσχερέστερον ἐπιλελογίσθαι κατὰ συνεγγισμὸν τῶν πρὸς τὸ ἀξιοπιστότερον εἰλημμένων θέσεων ἢ σχηματισμῶν, ἵνα μηδὲν ἡμῖν τῶν ἐνταχθησομένων εἰς συμπλήρωσιν τῆς ὅλης οἰκουμένης ἀόριστον ἔχῃ τὸν τόπον.” (...the numbers of degrees in longitude and latitude of well- trodden places are to be considered as quite close to the truth because more or less consistent accounts of them have been passed down without interruption; but [the coordinates] of the [places] that have not been so travelled, because of the sparseness and uncertainty of the research, have been estimated according to their proximity to the more trustworthily determined positions or relative configurations, so that none of the [places] that are to be included to make the oikumene complete will lack a defined position.) Trans. Berggren (2000) 94. 115/448 rather tedious lists of obscure toponyms related to sets of numeric data compounds the problem of an already unstable manuscript tradition. 379 Despite these problems, it is instructive to consider the spaces Ptolemy chose to work with, and how he divided the world as he saw it between them. It is difficult for us to relate Ptolemy’s six districts (τὰ χωρία) to the geography of Mesopotamia; it may have been difficult for Ptolemy as well. In his description of Syria Coele, he lists the various cities and villages according to the district in which they were located. In his description of Mesopotamia, he lists the included settlements only in relation to the the Tigris and Euphrates, while the districts of Mesopotamia are listed separately (Map 17: Ptolemy's Mesopotamia). Κατέχει δὲ τῆς χώρας τὰ μὲν πρὸς τῇ Ἀρμενίᾳ ἡ Ἀνθεμουσία, ὑφ’ ἣν ἡ Χαλκῖτις, ὑπὸ δὲ ταύτην ἥ τε Γαυζανῖτις, καὶ πρὸς τῷ Τίγριδι ποταμῷ ἡ Ἀκαβηνή, ὑπὸ δὲ τὴν Γαυζανῖτιν ἡ [Τ]ιγγηνή, καὶ ἐπὶ πολὺ παρὰ τὸν Εὐφράτην ἡ Ἀγκωβαρῖτις. 380 The location of Anthemusia, between the Euphrates and the Balikh around Batnae, is well-known; 381 however, it is not clear why Ptolemy should describe that region as “near Armenia” (πρὸς τῇ ʼΑρμενια). Dillemann identifies Gauzanitis (Γαυζανῖτις) with the region of Gauzan associated with the Khabur in II Kings, modern Ras el‘Aïn at the headwaters of the Khabur. 382 By analogy with that region and because Ptolemy places the area he calls Chalkitis (Χαλκῖτις) between Anthemusia and Gauzanitis, Dillemann suggests that Χαλκῖτις may be a corruption of Βαλικῖτις, and represent a region around the headwaters of the Balikh. 383 If this is the case, Ptolemy gives a progression of districts from west to east across the rain-fed northern belt of Mesopotamia: Anthemusia, Balikitis, Gauzanitis, and Akabene near the Tigris (καὶ πρὸς τῷ 379 Berggren (2000) 5, 41–50. 380 Ptol. Geog. 5.18.4: “[Mesopotamia] has these areas: the places near Armenia, Anthemusia; below which, Chalkitis; and below that: Gauzanitis; and towards the river Tigris, Axabene; and below Gauzanitis, [T]ingene [Singene?]; and then very close to the Euphrates, Ankobaritis.” 381 Appendix 1.3. 382 Dillemann (1962) 102; II Kings 17.6: “ἐν ἔτει ἐνάτῳ Ωσηε συνέλαβεν βασιλεὺς Ἀσσυρίων τὴν Σαμάρειαν καὶ ἀπῴκισεν τὸν Ισραηλ εἰς Ἀσσυρίους καὶ κατῴκισεν αὐτοὺς ἐν Αλαε καὶ ἐν Αβωρ ποταμοῖς Γωζαν καὶ ορη Μήδων.” (In the ninth year of Hoshea the king of Assyria seized Samaria, and deported Israel to Assyria, and settled them in Halah, and in Gozan on the Habor (Khabur) river, and in the land of the Medes.) 383 Dillemann (1962) 102. 116/448 Τίγριδι ποταμῷ ἡ Ἀκαβηνή). He then locates [T]ingene below Gauzanitis (ὑπὸ δὲ τὴν Γαυζανῖτιν ἡ [Τ]ιγγηνή). [T]ingene should be emended to Singene, the region around Singara. Finally , south of all of these regions, he places Ankobaritis along the northern bank of the Euphrates (καὶ ἐπὶ πολὺ παρὰ τὸν Εὐφράτην ἡ Ἀγκωβαρῖτις). Assuming that Χαλκῖτις is a mistake for Balikitis poses no conflict with Ptolemy’s coordinate locations because the name “Balikh” does not appear in Ptolemy’s description of Mesopotamia. He gives the location of two rivers flowing south from the foothills towards the Euphrates, the Khabur and the Saokoras 117/448 Map 17: Ptolemy's Mesopotamia (ὁ καλούμενος Σαοκόρας), but the Khabur is the most westerly of the two. It might be suggested that Ptolemy has reversed the positions of the two rivers and misnamed the Balikh as Saokoras, but the Balikh should flow close to Carrhae and emerge near Nikephorion on the Euphrates. Neither river conforms to that topography; Ptolemy places both Carrhae and Nikephorion to the west of the Khabur. Moreover, the Khabur is in more or less the correct place, while Ptolemy has the source of the Saokoras near Nisibis. Arab geographers identified the Saokoras as a tributary of the Khabur. 384 The confused relationship between the towns and rivers of Mesopotamia hints at the degree of error inherent in Ptolemy’s coordinates, his tradition, and ultimately the methodologies involved in the collection of his data. Repeated scientific observation of important locations should have been more accurate than estimations of physical movement, but our knowledge of the sources of Ptolemy’s data is insufficient to judge for any particular sites. Ptolemy gives coordinate locations for most of the settlements mentioned in the Mesopotamian descriptions of Strabo and Pliny , although not all of those settlements fall within Mesopotamia as Ptolemy defines it. Strabo discusses Zeugma in his descriptions of both Syria and Mesopotamia, while Ptolemy places it in Syria Coele (5.15.14, specifically in Cyrrhestice). Ptolemy places Tigranocerta in Armenia (5.13.22) and Thapsacus near the Euphrates in Arabia Deserta (5.19.3) rather than Mesopotamia. Ptolemy does not mention the toponyms Sinnaca or Chordiraza, 385 but he places Nikephorion near the Euphrates and Nisibis and Carrhae in the middle region. 386 Pliny referred to Zeugma, Edessa, Carrhae, Anthemusia and Nicephorium in book five; Ptolemy has Edessa in the middle region but he mentions Anthemusia only as a district, omitting the town of Batnae/Anthemus. In book six, Pliny referred to Nisibis as Antioch (a dynastic name also given by Strabo), Ptolemy has Nisibis only . Of the two short lists of towns given by Pliny 384 Musil, (1927) 339–40, discusses how the Arabic geographers who used Ptolemy reconciled the identification of the Saokoras and a tributary of the Khabur called el-Hermâs. 385 Ptolemy includes two settlements called Sinna (Σίννα, 5.18.11 & 12), but both are too far from Carrhae to be linked to the hill called Sinnaca, see Appendix 1.17. Dillemann relates Chordiraza to Βαρσάμψη (5.18.5), see Appendix 1.7. 386 Νικηφόριον (5.18.6), Νίσιβις (5.18.11) and Κάῤῥαι (5.18.12). 118/448 (Seleucia, Laodicea, Artemita; and Dios Pege, Polytelia, Stratonicea), only Seleucia on the Tigris appears in Ptolemy’s list of Mesopotamian communities, although he does locate towns called Artemita in Arabia Deserta and Assyria. 387 Dillemann suggests that Oruros may be listed in Ptolemy as Orthaga (Zagura). 388 Neither of Caphrena or Thebata appear in Ptolemy’s lists. Adjacent Areas In his description of Mesopotamia, Ptolemy gives Mesopotamia five edges. A brief consideration of the adjacent regions which form those edges reveals a few details about the delimitations of the region in Ptolemy’s account. Three of those regions (Syria, Arabia Deserta, and Assyria) are explicitly defined by their relationship to the Euphrates and Tigris rivers. 389 Syria Coele comprises all of northern Syria south of the Amanus and from the Euphrates to the sea. 390 Ptolemy divides Syria Coele into districts, then locates the cities of Syria within those districts. Whereas in Mesopotamia, Ptolemy did not relate his verbal description of the districts to his coordinate lists of the cities and villages, in the Syrian chapter, the two are interwoven, with the cities listed according to the district in which they are found. This suggests that Ptolemy was better informed about Syria than Mesopotamia. Moreover, he is able to give relative position within those districts; within several of his subdivisions of Syria, Ptolemy separately lists cities along the Euphrates in Commagene (5.15.10-11), Cyrrhestice (5.15.13-14), Chalybonitis (5.15.17), and Palmyrene (5.15.24-25). These districts, as well as Chalkidice (5.15.18), which he does not extend as far as the Euphrates, are the areas relevant to the Mesopotamian borderland. 387 Σελεύκεια πόλις (5.18.8); Ἀρτέμιτα (5.19.7 & 6.1.6). 388 Ptol. Geog. 5.18.12. Dillemann (1962) 311. See above, p.74. 389 Ptol. Geog. 15.18.1. Syria and Arabia Deserta by the Euphrates to the west and south respectively and Assyria by the Tigris to the east. 390 Ptol. Geog. 5.15.7. 119/448 There are few discrepancies between Ptolemy’s location data and that of Strabo and Pliny; however, neither of those authors mention the presence of Chalybonitis. It can be safely located on the west side of the Euphrates bend from Ptolemy’s inclusion of Barbalissos among those cities along the Euphrates. Based on Ptolemy’s coordinates, it lay east of Chalkidice and may correspond to part of the region of pastoralists that Strabo called Parapotamia. 391 Pliny also places a group of Skenitai Arabs in this area. 392 Ptolemy’s account may indicate a greater degree of urbanisation in this area, or show a gap in the knowledge or attention of 391 Strabo 16.2.11. 392 Pliny NH 5.87. 120/448 Map 18: Ptolemy's Syrian Districts those earlier authors. It is interesting to note that it is while describing precisely this area that Strabo remarks that the pastoralists on the edges of Syria are more civilised the closer they dwell to the (settled and urbanised) Syrians. 393 Ptolemy’s Palmyrene also includes some cities on the Euphrates which would have fallen within this area within which Strabo describes only pastoralists. Moreover, his description of Palmyrene is far more populated than Pliny’s account indicated; Ptolemy gives coordinates for 16 cities in all directions around the city which Pliny took pains to isolate. While Syria Coele faces Mesopotamia across the Euphrates predominantly to the east, Arabia Deserta extends south from that river after the district of Palmyrene. It is notable how many cities Ptolemy locates in this region: three near the Persians, 26 in the interior and ten along the Euphrates, including Thapsacus. 394 Across the Tigris, Ptolemy defines Assyria by the adjacent regions of Greater Armenia and the Niphates mountains to the north, Media to the east, Susiana to the south, and the Tigris to the west. 395 Similarly to his treatment of Mesopotamia, Ptolemy describes the internal subdivisions verbally and without coordinate locations, then organises the region’s coordinate locations into two groups according to proximity to the river or position in the interior. 396 Ptolemy’s Assyria includes the entire plain east of the Tigris and below the Niphates and Zagros mountains, including the plains north of Nineveh and as far south as Ctesiphon and Sittacene. 397 The northern boundary of Ptolemy’s Babylonia, that with Mesopotamia, is further south than we 393 Strabo 16.2.11. 394 For the implications of Ptolemy’s location of Thapsacus, see Appendix 1.18. 395 Ptol. Geog. 6.1.1: “Ἡ Ἀσσυρία περιορίζεται ἀπὸ μὲν ἄρκτων τῷ εἰρημένῳ τῆς Μεγάλης Ἀρμενίας μέρει παρὰ τὸν Νιφάτην τὸ ὄρος, ἀπὸ δὲ δύσεως Μεσοποταμίᾳ κατὰ τὸ ἐκτεθειμένον τοῦ Τίγριδος ποταμοῦ μέρος, ἀπὸ δὲ μεσημβρίας Σουσιανῇ κατὰ τὴν ἀπὸ τοῦ Τίγριδος γραμμὴν μέχρι πέρατος...” (Assyria is bordered to the north by the part of Greater Armenia neat the Niphates Mountains, to the west by the part of Mesopotamia near the Tigris river, to the south by Susiana right up to the far line of the Tigris...) 396 Ptol. Geog. 6.1.2. Ptolemy’s arrangement of Adiabene is discussed by Marciak (2011) 194–95. 397 Νῖνος and Κτησιφών (6.1.3). Among the districts, Sittacene lies near Susiana (“ἡ δὲ παρὰ τὴν Σουσιανὴν Σιττακηνή”, 6.1.2), previous defined as the southern boundary (6.1.1). Kalachene (ἡ Καλακινή, 6.1.2) lay north of Nineveh, see above p.89. 121/448 might expect. 398 Ptolemy included the two most politically important cities of Babylonia in the regions to the north; Seleucia on the Tigris in Mesopotamia (5.18.8) and Ctesiphon in Assyria (6.1.3). Ptolemy locates the cities of Babylonia in five areas, the two northernmost of which are those near the Euphrates and those along the river which flows through Babylon. 399 In that second set, Ptolemy includes Babylon, the canal Naarmalcha, Volgaises, and Barsita. 400 This corresponds to the boundary that Ptolemy drew between Mesopotamia and Babylonia in his description of the former: namely , that it ran along “the connection to the Tigris” (τῆς πρὸς τὸν Τίγριν συναφῆς). Ptolemy draws the northern boundary of Babylonia along the most convenient water boundary , the Naarmalcha canal which connects the Euphrates to the Tigris. He placed the cities that lay north of that water course in Mesopotamia and those to the south in Babylonia. Similarly , he places the cities on the east side of the Tigris (like Ctesiphon) in Assyria rather than Babylonia. Where water bodies were available, Ptolemy preferred to define space according to their courses, and he did so strictly . To the north, however, there was no convenient river by which to delimit Mesopotamia. Ptolemy defines the boundary between Greater Armenia and Mesopotamia according to the Taurus, between the places through which the Euphrates and Tigris rivers cut. 401 As well as those points, he gives coordinate information for the sources of both rivers in Armenia and notes that the Tigris flows through lake Thospitis (“λίμνην τὴν καλουμένην Θωσπῖτιν”). 402 Ptolemy uses those sources and lake Thopsitis to locate 398 Ptol. Geog. 5.20.1: “Ἡ Βαβυλωνία περιορίζεται ἀπὸ μὲν ἄρκτων Μεσοποταμίᾳ κατὰ τὸ ἐκτεθειμένον τοῦ Εὐφράτου μέρος, ἀπὸ δὲ δύσεως τῇ Ἐρήμῳ Ἀραβίᾳ κατὰ τὴν εἰρημένην ὀρεινὴν, ἀπὸ δὲ ἀνατολῶν Σουσιανῇ παρὰ τὸ λοιπὸν τοῦ Τίγριδος μέρος μέχρι τῶν εἰς τὸν Περσικὸν κόλπον ἐκβολῶν...” (In the north, Babylonia borders on that part of Mesopotamia next to the Euphrates, to the west Arabia desert near the desolate hills, to the east Susiana near the remaining part of the Tigris as it flows out into the Persian gulf.) 399 Ptol. Geog. 5.20.6: “Παρὰ δὲ τὸ τοῦ Εὐφράτου ποταμοῦ μέρος αἵδε... καὶ ἐπὶ μὲν τοῦ διὰ Βαβυλῶνος ῥέοντος ποταμοῦ...” ([The cities and villages] in the part near the Euphrates River are... and on the river that flows through Babylon...) 400 Ptol. Geog. 5.20.6: Βαβυλών, πρὸς δὲ τῷ Μααρσάρῃ ποταμῷ, Οὐολγαισία, Βάρσιτα. For Marses as a corruption of Naarsares/Nar Sarri (Naarmalcha), canal of the king, see Appendix 1.12. The first set (Παρὰ τὸ τοῦ Εὐφράτου ποταμοῦ μέρος) comprises Ἰδικάρα, Δούραβα, Θάκκονα, Θελβεγκάνη. 401 Ptol. Geog. 5.13.4. 402 Ptol. Geog. 5.13.7. 122/448 the district of Anzetene (“Ἀνζητηνὴ (ἢ Ἀνζιτηνὴ)”). 403 If the identification of Lake Van with Thospitis is correct, then Ptolemy is describing the course of the Botan, not the upper Tigris; just as Strabo did. 404 This would make Anzetene a region on the southern slopes of the Taurus, north of the Tigris and west of the Botan. Ptolemy places Gordyene in Armenia east of the source of the “Tigris”, and locates Tigranocerta within Gordyene. 405 This placement of Gordyene on the southeast side of the Botan and east of the Tigris proper is consistent with earlier classical sources like Strabo and Xenophon. 406 It is clear, especially from the boundaries that Ptolemy assigns, that Ptolemy’s conception of the Mesopotamian borderland was predominantly geographical. His geography attempts to construct a rigorous and objective account of space by precise quantification of every relevant geographical feature for which he had information. He makes no mention of the political formations which occupied the spaces he describes. Those political formations are only evident indirectly . By comparing his work with modern geographical, archaeological and historical knowledge, we can see that his knowledge becomes weaker as his text moves further from the Roman empire. 407 The Expositio Totius Mundi et Gentium The Expositio begins by describing tribes far to the east of the Roman empire, eventually reaching in turn the Persians, the Saraceni and Roman territory . It then proceeds province by province around the Empire, briefly describing the province, its inhabitants and one or two of its major cities. In this, the author fulfils his aim of describing all the barbarian nations and Roman provinces, and the cities and outstanding 403 Ptol. Geog. 5.13.18: “Ἐν δὲ τῷ λοιπῷ καὶ μεσημβρινωτέρῳ τμήματι μεταξὺ μὲν τοῦ Εὐφράτου καὶ τῶν τοῦ Τίγριδος πηγῶν ἥ τε Ἀνζητηνὴ (ἢ Ἀνζιτηνὴ) καὶ ὑπ’ αὐτὴν ἡ Θωσπῖτις, εἶτα Κοριναία.” 404 See above, p.89. 405 Gordyene definitely lay within Armenia (Ptol. Geog. 5.13.20) but the line which describes the section of coordinates including Tigranocerta as being within Gordyene is bracketed without explanation in Nobbe’s edition: Ptol. Geog. 5.13.22: [καὶ ὑπ’ αὐτὴν ἡ Γ ορδυνη[σία] ἢ Γ ορδυηνή ....... οε δʹ λθ γʹ], Nobbe (1966) 2.55. 406 Although note that Strabo, who takes a more ethnographic view than Ptolemy , has the lands of the Gordyeans extending on both sides of the Tigris south (downstream) of its confluence with the Botan. See above, p.89. 407 For Ptolemy on India, see Parker (2008) 74–76. 123/448 features of the latter. 408 There is some variation between the areas beyond the frontier and those within it. Those beyond the frontier comprise a linear narration according to ethnic categories, while those within Roman space are organised according to the administrative structure of the Roman Empire. In both parts, the work is unconcerned with topographical information. The Expositio Totius Mundi treats geographical space beyond the bounds of the Roman Empire as a simple list of areas. At the start of this part of the Expositio, the author indicates that the general direction of description is westerly (“ad occidentem”). 409 Thereafter, he gives no directional information, instead connecting each brief description with a simple conjunction (deinde) or conjunctive phrase with post. 410 Each description ends with a practical measurement of the extent of that territory in terms of the number of staging posts (mansiones) that a traveller would pass in crossing it. The only other topological information is in the use of adiacio (to border on) to indicate territorial adjacency . 411 This style gives the impression of a connected line of areas rather like neighbourhoods on a bus or train route. 412 The impression of a linear progression is strengthened by the use of sequor when the narration reaches Roman territory . 413 The territories linked in this way are described in either geographic or cultural terms. Where a reader might not be expected to know the people or land in question, regio or gens is used with the proper name, 408 Expositio 2: “Quaerentes autem scribere, debemus dicere primum [quando mundus a deo fuerit institutus dehinc] quae gentes ab oriente usque ad occidentem constitutae sint; post hoc quanta sint genera barbarorum, deinde omnem Romanorum terram, quot sint in omni mundo provinciae, vel quales in substantia ac potestate; quae civitates in singulis provincis habeant et quid in unaquaque provincia aut civitate possit esse praecipuum. Munificum enim hoc opus et studiosum mihi esse videtur” (But seeking to write, we ought to say first […] which nations have been established from the east to the west; after this, how many kinds of barbarians, then the whole land of the Romans, how many provinces there are in the whole world, or what kind of wealth and power they have; which cities are in each province and what is outstanding in every province and city. For this work seems to me to be liberal and studious.) The concerns of the Expositio are primarily commercial and will be discussed in Chapter 5. 409 Expositio 8. 410 Post: Expositio 8, 9, 10, 11, 13, 14, 18, 19, 20, 22. Deinde: Exposito 15, 16, 17. 411 Expositio 13, 17. 412 Also noted by W oodman (1964) 52–53. 413 Expositio 22: “Post hos nostra terra est. Sequitur enim Mesopotamia et Osdroena.” (After this is our land. Mesopotamia and Osrhoena follow . 124/448 but where knowledge of certain lands or peoples is assumed, the proper name stands alone. 414 As expected, the latter cases increase in frequency as the narrative draws closer to Roman space. This suggests that the author expects that a significant proportion of his audience will not be familiar with these extra-Roman territories. 415 For the author of the Expositio, space beyond the Roman Empire is imagined as a linear path through a series of geographic or cultural spaces, an itinerary , as a passing foreign traveller or trader might experience it. The source of this information is unclear, but it is not first hand. Just after his description of those extra-Roman territories, the author notes that his source is an unnamed historicus. 416 This marks a break between information explicitly gathered from an unnamed source (the historicus), and information over which the Expositio implicitly claims a more expert knowledge. As the Expositio begins to describe Roman space, the author re-asserts his authority over his material with a first person statement of intent. 417 The greater depth of information and complexity of topological information in the descriptions which follow support this implicit claim, although not without error. In fact the Expositio is only concerned about topological information in a very restricted sense. As can be seen in the description of Mesopotamia et Osdroena, the text identifies the Roman province (or group of provinces) which it is about to discuss, then provides little more than a list of cities within it. 418 414 Regio or gens with proper name: Expositio 9 (regio Eviltarum), 10 (gens quae vocatur Emer), 11 (regio quae appellatus Nebus), 13 (regio Disaph), 14 (gens sic appellanda Choneum), 15 (Diva gens), 17 (Exomia regio), 22 (Sarracenorum gens). Proper name alone: Expositio 16 (India maior), 18 (India minor), 19 (Persae, Romanis propinquantes), 22 (Mesopotamia et Osdroena). 415 On the place of these eastern locations in the Roman imagination, see Parker (2008). 416 Expositio, 21: “et haec quidem de praedictis gentibus historicus ait.” (And in fact, an historian says this about the preceding peoples). 417 Expositio, 21: “Quoniam vero necessarium est et nostram terram, hoc est Romanorum, conscribere, experiar exponere, ut possit legentibus prodesse. Incipiamus ergo” (Since truly it is necessary to survey our land, that is, the land of the Romans, I will attempt to explain so that it can be useful to readers. So we will begin.) 418 Expositio 22: “Post hos nostra terra est. Sequitur enim Mesopotamia et Osdroena. Mesopotamia quidem habet civitates multas et varias, quarum excellentes sunt quas volo dicere. Sunt ergo Nisibis et <Amida>, quae in omnibus viros habent optimos et in negotio valde acutos et bene venantes. Praecipue et divites et omnibus bonis ornati sunt: accipientes enim a Persis ipsi in omnem terram Romanorum vendentes et ementes iterum tradunt, extra aeramen et ferrum, quia non licet hostibus dare aeramen aut ferrum. Istae autem civitates semper stantes deorum et imperatoris sapientia, habentes moenia inclita, bello semper virtutem Persarum dissolvunt; ferventes negotiis et tranigentes cum omni provincia bene. Deinde Osdroenae Edessa et ipsa civitas splendida.” (After this 125/448 Sometimes topological relations are implied, such as a reference to a port or naval trade placing a city on a coastline. In Mesopotamia, the Expositio refers to only two cities (Nisibis and Amida) 419 ; in Osdroena (Osrhoena), only one, Edessa. The spaces of the two provinces are not delimited individually or as the unit into which the text aggregates them, instead it is left to the reader to identity the names which the text provides as Roman provinces and then to supplement that information with their own knowledge as required. This method is even more clear in the Expositio’s more detailed description of Syria. After Mesopotamia and Osrhene, the Expositio moves on describe the three Syrian provinces, Syria Punica, Syria Palestina, and Syria Coela, together and at some length. 420 Besides making this initial distinction, the Expositio treats all three provinces as one area and shows no concern for accurate topological description. The narrative jumps around Syria from city to city , connected by conjunctive adverbs like iam 421 and post 422 without adhering to a topological order. Instead, the narrative highlights large, important, coastal cities and includes inland and smaller coastal cities in lists. 423 These cities are described almost exclusively in economic terms, except for the section on Antioch’s regional primacy , and the importance of the legal school at Berytus. 424 is our land; for Mesopotamia and Osrhene follow . Indeed, Mesopotamia has many diverse cities; I wish to speak of those which are outstanding. Namely , Nisibis and <Amida>, which have men who are the best men in all things, very acute in business, and good salesmen. The cities are especially wealthy and supplied with all goods: for they receive sellers from Persia into the whole Roman empire and send back buyers. Except bronze and iron, since it is not permitted to give bronze or iron to foreigners. But those cities ever-standing by means of the forethought of the gods and the emperor, having famous walls, always destroy the courage of the Persians in war. Fervent in business and dealing well with every province. Then Edessa of Osdroenae also a very splendid city.) 419 The manuscript’s first reference to Edessa should be emended to Amida. See Chapter 5. 420 Expositio, 23-33. 421 Expositio, 24, 26, 27 422 Expositio, 25 (accepting the textual reconstruction from the Descriptio, “<Post istam> Berytus...”) 423 Highlighted cities: Antioch (Expositio 23) and her two ports Laodicea (27) and Seleucia (28), Tyre (24), Berytus (25), Caesarea (26), and Ascalon and Gaza (together at 29). Listed cities: Neapolis, Tripolis, Scythopolis, Byblus, Heliopolis, Sidon, Sarepta, Ptolemais, Eleutheroplis, Damascus (Expositio 30). While Heliopolis receives a note about the beauty of its women, it is otherwise undescribed and appears between two lists of cities. 424 Economic descriptions of Tyre (Expositio 24), Caesarea (26), Ascalon and Gaza (29); the political importance of Antioch (23); the legal school at Berytus (25). After the Expoitio’s winding survey of Syria, the text includes a specific section on produce and trade goods in Syria: Expositio 31. 126/448 The Expositio uses categories which could be defined geographically (provinces and regions) but does it not define them in this way itself. It relies on the reader’s prior knowledge of geography and applies the information about which it is concerned atop that knowledge. It does not attempt to organise space, so much as to make qualitative statements about features within it. The narrative’s linear progression towards the Empire from the east is a sort of ethnic itinerary of non-Roman space which only places those peoples in a linear relation to each other. Within Roman space the provinces (individually or aggregated) act as analytical containers for the cities about which the text is primarily concerned. That commercial, political, and ideological information which the Expositio attached to those cities will be discussed in subsequent chapters. Ammianus Marcellinus Ammianus’ history includes several geographical digressions or sequences in his work. The three which this section examines are those of importance for his conception of the Mesopotamian borderland. The first is his survey of the eastern provinces of the Roman Empire. This discussion is necessarily short owing to the loss of Ammianus’ description of the Roman provinces of Mesopotamia and Osrhoena itself; however, it provides a opportunity to examine Ammianus’ credentials as a geographical source. The second geographical section is his excursus on the regions of the Persian empire, especially his account of “Assyria”. That region included Adiabene and southern Mesopotamia, both of which he describes largely in terms of the cities within them. That second geographical sequence falls within the third, Ammianus’ description of Julian’s invasion of the Persian region of Assyria in 363 BCE. This is the most complete account of a military expedition along the Euphrates route. It contains a number of geographical details concerning cities, towns and fortresses along the routes to and from Ctesiphon, but is most important for what it tells us about the borders between Roman and Sasanian space both along the Euphrates and near Nisibis. 127/448 The Eastern Provinces Near the start of the extant portion of his work, Ammianus describes the orientales provinciae, beginning with Cilicia and Isauria and ending with Arabia. 425 Unfortunately , Ammianus had already described Mesopotamia in an earlier book, now lost, and he declines to repeat or elaborate on that description. 426 Here, Ammianus is referring to “Mesopotamia” as a geographical region, presumably including the Roman province of Mesopotamia, but also including the province of Osdroena. 427 The omission of these provinces from his survey is particularly painful given Ammianus’ first hand experience of the region and the quality of his other geographical descriptions. 428 Ammianus displayed his erudition through references to and citations of important works and figures in the historiographical and geographical traditions that preceded him, but he also relied on geographical autopsy to a degree uncharacteristic of those traditions. 429 For the geography of the eastern provinces, Ammianus’ autopsy derived from his upbringing in Antioch and his military service in Mesopotamia. That lost description of Mesopotamia and Osdroena occurred in a context describing the Parthian wars. Ammianus’ history focused on the Roman emperors from Nerva (96-98 CE) to Valens (378 CE) and their activities at court, in civil conflicts and in their foreign expeditions. 430 The historical context of the lost description was probably an account of Trajan’s 425 Amm. Marc 14.8. Egypt would be included, but Ammianus defers that description for later in his work (it occurs at Amm. Marc. 22.15-16), Amm. Marc. 14.7.21. 426 Amm. Marc, 14.7.21: “...absque Mesopotamia, iam digesta cum bella Parthica narrarentur...” (...except Mesopotamia which has already been described along with the Parthian wars...) 427 Amm. Marc. 14.8.7: “Et prima post Osdroenam quam, ut dictum est, ab hac descriptione discrevimus, Commagena (nunc Euphratensis), clementer assurgit, Hierapoli (vetere Nino) et Samosata civitatibus amplis illustris.” (And first after Osdroene, which, as has been said, I have omitted from this account, Commagene, now called Euphratensis, gradually rises, famed for the great cities of Hierapolis (ancient Ninus) and Samosata.) 428 In particular, his detailed descriptions of Gaul (15.10-11), Lower Pannonia (21.10.2-4?), Egypt (22.15-16) and Thrace (27.4). 429 Sundwall (1996) 624–25. 430 Amm. Marc. 31.16.9: “Haec ut miles quondam et Graecus, a principatu Caesaris Nervae exorsus ad usque Valentis interitum pro virium explicavi mensura: opus veritatem professum numquam, ut arbitror, sciens silentio ausus corrumpere vel mendacio. scribant reliqua potiores, aetate doctrinisque florentes. quos id, si libuerit, adgressuros, procudere linguas ad maiores moneo stilos.” (I, as a former soldier and a Greek, have laid out this work to the best of my ability , beginning from the of the reign of Nerva up to the death of Valens. It professes to be the truth, as I bear witness, never knowingly daring to corrupt it by silence or a lie. Let more able men, in the bloom of age and learning, write the rest. I advice those who will undertake the task, if desired, to produce 128/448 Mesopotamian campaign, but if not then, he would have found ample occasion for a detailed examination of the space in relation to the campaigns of Lucius Verus or the Severans during which the Roman provincial structures of Mesopotamia and Osrhoene were established. There is no evidence in the surviving books of any attempt at a systematic view of the empire; Ammianus’ gaze does not fall on a region except to set the scene for some sort of conflict. As his period excludes the period during which Syria was incorporated into the empire, he probably did not have cause to discuss its geography in any detail beyond this section on the eastern provinces. Nevertheless, it is important to note that Ammianus usually delimited space according to political and administrative divisions, as would be usual to a man employed by the imperial administration. The Persian Regiones The second of Ammianus’ provincial surveys occurs at the beginning of his account of Julian’s expedition against southern Mesopotamia in 363 CE. 431 Suitably for such a context, Ammianus begins with a short historical summary of the history of the Mesopotamian borderland before beginning with the locorum situm (arrangement of places) itself. 432 He begins by describing the Persian Gulf and attempting to situate the Persian empire relative to it. Utque geographici stili formarunt, hac specie distinguitur omnis circuitus ante dictus. Ab arctoo cardine usque ad Caspias portas, Cadusiis conterminat et Scytharum gentibus multis et Arimaspis hominibus luscis et feris. Ab occidua plaga contingit Armenios et Niphaten et in Asia sitos Albanos, Mare Rubrum et Scenitas Arabas quos Saracenos posteritas appellavit: Mesopotamiam sub axe meridiali despectat: orienti a fronte contrarius ad Gangen extenditur flumen, quod Indorum intersecans terras, in pelagus eiectatur australe. 433 language of the ancient style.) 431 Amm. Marc. 23.6.1-74. 432 History of imperial conflicts in the borderland: Amm. Marc. 23.6.2-9. Amm. Marc. 23.6.10: “Nunc locorum situm (quantum ratio sinit), carptim breviterque absolvam.” 433 Amm. Marc. 23.6.13: “And as the pens of geographers have formed it, the whole circuit just described is given this form. In the northern direction, to the Caspian Gates it borders on the Cadusii, on many tribes of the Scythians, and on the Arimaspse, wild, one-eyed men. On the west it touches Armenia, Niphates, the Asiatic Albani, the Red Sea, and the Skenitai Arabs, whom men of later times called the Saracens. It looks down on Mesopotamia under the southern sky. Opposite the eastern face it extends to the Ganges river, which cuts through the lands of the Indians and empties into the southern ocean.” 129/448 The first clause of this passage can be read as evidence that Ammianus had knowledge of maps of the Persian empire, but as with many such potential literary references to maps, the words are ambiguous. 434 The pens of the geographers (geographici stili) could form or shape (formarunt) a literary narrative, although the most natural use of formare would denote an image or shape of some sort. Forma was the usual word used to denote a map or cadastral plan. 435 Globes were certainly known to Strabo and Ptolemy , so there is no problem in allowing Ammianus access to one. The description that follows suggests a deficiency in the image or in Ammianus’ memory or interpretation of it. The circuit (circuitus) Ammianus is referring to as “just described” (ante dictus) comprises the regions (regiones) of the Persian Empire as they are arranged around the Persian Gulf (see Map 19: Ammianus' Ciruitus). To the north of this circuit, Ammianus places the Caspian Gates, the Cadusians, the Scythians and the mythical Arimaspse; 436 to the west he has Armenia, Niphates, the Asiatic Albani, the Red Sea, and the Skenitai Arabs; 437 to the east lies the Ganges and India. However, he then places Mesopotamia to the south (Mesopotamiam sub axe meridiali despectat “it looks down on Mesopotamia under the southern sky”). The passage purports to give a description of the lands around the Persian gulf, so we might expect that gulf to be the centre from which the narrative looks out, but this is not the case. Instead the spatial arrangement described is that of the Persian empire itself. The best fit for an imaginary central viewing point that has the Caspian Gates, the Cadusians, the Scythians and the mythical Arimaspse to the north, Armenia, Niphates, the Asiatic Albani, the Red Sea, and the Skenitai Arabs to the west, the Ganges and India to the east, and Mesopotamia (alone!) to the south, is somewhere in Media or the Zagros mountains. This hints that Ammianus’ description of the Persian Empire relied on a source composed in 434 Sundwall (1996) 622. 435 Dilke (1985) esp. 196. 436 The Cadusians were an Iranian people dwelling south of the Caspian Sea, not to be confused with the Cardusians/Gordyaeans: Strabo 11.8.1; von Bredow , Iris ‘Cadusii’ BNP . Arimaspse: Hdt. 4.13.1. 437 For the Niphates, see p.40. The Albani dwelt in the Caucasus region, north of Armenia: von Bredow , Iris ‘Albani’ BNP . 130/448 Media, perhaps at Ecbatana. Indeed, the description never identifies any space as Roman, which may hint that the ultimate source was composed in the Persian Empire. 438 The description of the regiones proper begins with a list of those regions, beginning at Assyria and proceeding more or less clockwise around the empire. Sunt autem in omini Perside, hae regiones maximae, quas vitaxae (id est magistri equitum) curant, et reges et satrapae—nam minores plurimas recensere difficile est et superfluum—Assyria, Susiana, Media, Persis, Parthia, Carmania maior, Hyrcania, Margiana, Bactriani, Sogdiani, Sacae, Scythia infra Imaum et ultra eundem montem, Serica, Aria, Paropanisadae, Drangiana, Arachosia et Gedrosia. 439 438 On the composition and sources of Ammianus’ history , see Matthews (1989) 17–32; Kelly (2008) 222–55. While Ammianus is consistent in his use of the name “Mesopotamia” to mean the Roman province, it is possible that if this reference comes from a Persian source, a geographical term which includes part or all of Babylonia may have been translated as Mesopotamia and included by Ammianus. 439 Amm. Marc. 23.6.14: “However there are in all Persia these greatest regions, administered by vitaxae (which are 131/448 Map 19: Ammianus' Ciruitus These regions are explicitly administrative; each is under the care of a vitaxa (glossed as a magister equitum, quas vitaxae (id est magistri equitum) curant). 440 Ammianus follows the clockwise arrangement set out by this introductory passage in the geographical narrative that follows, with a passage on Arabia Felix inserted between those describing Parthia and Greater Carmania. According to this division of space, Ammianus’ section on Assyria describes those parts of the Mesopotamian plain controlled by Sasanid Persia and adjacent to Roman Mesopotamia including not just parts of northern Mesopotamia, but Adiabene and Babylonia. As the region containing the major zone of contact between the Roman and Sasanid empires, it is appropriate that the description of Assyria is the largest and most detailed part of Ammianus’ Persian geography . 441 Ammianus’ use of “Assyria” to refer to a wide swathe of the fertile crescent is reminiscent of Strabo’s understanding of the term. The important difference between the two is that Strabo includes Roman Syria within Assyrian lands, whereas Ammianus only includes non-Roman lands. However, there may be a hint of Strabo’s use in Ammianus’ initial description: Citra omnes provincias est nobilis Assyria celebritate et magnitudine et multiformi feracitate ditissima. quae per populos pagosque amplos diffusa quondam et copiosa, ad unum concessit vocabulum et nunc omnis appellatur Assyria, ubi inter bacarum vulgariumque abundantiam frugum bitumen nascitur prope lacum nomine Sosingiten, cuius alveo Tigris voratus fluensque subterraneus percursis spatiis longis emergit. 442 The contrast Ammianus draws between the region once having been spread through many people and districts (quondam per populos pagosque amplos diffusa) and now being united under the single name of commanders of cavalry), and kings and satraps – for it is difficult and unneccessary to count the large number of lesser divisions – Assyria, Susiana, Media, Persis, Parthia, Greater Carmania, Hyrcania, Margiana, the Bactriani, the Sogdiani, the Sacae, Scythia at the foot of Imaus [the Himalayas] and beyond the same mountain, Serica, Aria, the Paropanisadae, Drangiana, Arachosia, and Gedrosia.” 440 A vitaxa (Persian Bidaxš, Greek βιτάχης, πιτιαχης, Armenian bdeašχ) is a toparch or marcher-lord, Sunderman, “Bidaxš”, EIran 4.242-44; Garsoïan (1989) 516–17; Shahîd (1984) 119–20, n.34. 441 Amm. Marc. 23.6.15-25. 442 Amm. Marc. 23.6.15: “The nearest to us of all the provinces is Assyria, famous for its large population, its size, and the great wealth and fruitfulness of its many products. This province once spread over great and prosperous peoples and districts, it submitted to a single name, and now the entire region is called Assyria. There among a great abundance of berries and common fruits, bitumen is found near the lake called Sosingites, in whose bed the Tigris is swallowed up, and then, after flowing under ground, emerges after traversing a long distance.” 132/448 “Assyria” (ad unum concessit vocabulum et nunc omnis appellatur Assyria) is primarily a contrast between a space inhabited by diverse populations that has become a unified space under a single name, but the use of diffusa per and concedo hints at a wider scope (such as the “Assyrian” space described by Strabo) that has been constricted in a smaller space (as Ammianus describes). The first specific area within Assyria that Ammianus describes is Adiabene, which he notes was once called Assyria (“Assyria priscis temporibus vocitata”). 443 Ammianus relates the Greek etymological theory that Adiabene was named from διαβαίνειν, to cross, because it was positioned between two rivers. 444 He refers to unnamed ancient authorities (“veteres quidem hoc arbitrantur”) that the two rivers were the Ona and Tigris, but suggests, based on his personal experience of the area, that the two rivers were in fact the Diabas and Adiabas. 445 These were probably the Greater and Lesser Zab. 446 While this etymological discussion provides the incidental geographical information that these two rivers formed the boundaries of Adiabene, Ammianus gives no further description of the physical geography of the region. As is often the case in his geographical descriptions, Ammianus is more concerned with the cities to be found in a region than with physical geography and topology . 447 His placement of Ninus (Nineveh), Arbela and Gaugamela in Adiabene is 443 Amm. Marc. 23.6.20. 444 In fact, it is from the Aramaic name for the region Ḥadyab. 445 Amm. Marc. 23.6.20-21: “Intra hunc circuitum Adiabena est, Assyria priscis temporibus vocitata longaque adsuetudine ad hoc translata vocabulum ea re quod inter Onam et Tigridem sita navigeros fluvios adiri vado numquam potuit: transire enim diabainein dicimus Graeci. [21] et veteres quidem hoc arbitrantur. nos autem id dicimus quod in his terris amnes sunt duo perpetui, quos ipsi transivimus, Diabas et Adiabas iunctis navalibus pontibus, ideoque intellegi Adiabenam cognominatam ut a fluminibus maximis Aegyptos Homero auctore et India et Euphratensis ante hoc Commagena, itidemque Hiberia ex Hibero, nunc Hispania, et a Baeti amne insigni provincia Baetica.” (Within this area is Adiabena, called Assyria in ancient times, but by long custom changed to this name because, lying between the navigable rivers Ona and Tigris it could never be approached by a ford; for we Greeks for transire say διαβαίνειν. At least, this is the opinion of the ancients. [21] But I myself say that there are two perpetually flowing rivers to be found in these lands, the Diabas and Adiabas, which I myself have crossed, and over which there are bridges of boats; and therefore it is to be assumed that Adiabena was named from them, as from great rivers Egypt was named, according to Homer, as well as India, and the Euphratensis, before my time called Commagena; likewise from the Hiberus, Hiberia (now Hispania), and the province of Baetica from the noble river Baetis.) 446 Marciak (2011) 196–99. 447 Amm. Marc. 23.6.22: “In hac Adiabena Ninus est civitas, quae olim Persidis regna possederat, nomen Nini potentissimi quondam regis Samiramidis mariti declarans, et Ecbatana et Arbela et Gaugamela, ubi Dareum Alexander post discrimina varia proeliorum incitato Marte prostravit.” (In this Adiabena is the city of Ninus, 133/448 entirely consistent with other Roman geographic writers. 448 He erroneously locates Ecbatana there as well, but elsewhere correctly places this city in Media. 449 The rest of Ammianus’ description of Assyria refers to places in southern Mesopotamia. He begins with a simple list of cities: “Apamea, formerly called Mesene, and Teredon, Apollonia and Vologessia”. 450 which once held sovereignty over Persia, proclaiming the name of Ninus, a once most powerful king and the husband of Semiramis; also Ecbatana, Arbela, and Gaugamela, where Alexander, after various other battles, overthrew Darius by the incitement of Mars.) 448 Ammianus elsewhere (18.7.1) describes Nineveh as the great city of Adiabene (“Nineve Adiabenae ingenti civitate”), 449 Amm. Marc. 23.6.39; Hdt. 1.98; Ecbatana is modern Hamadan. Calmeyer, Peter, ‘Ecbatana’ BNP . 450 Amm. Marc. 23.6.23: “In omni autem Assyria multae sunt urbes. inter quas Apamia eminet Mesene 134/448 Map 20: Adiabene in Context Vologesia (or Vologesocerta/Balashkert) was a Parthian city founded in the first century CE by Vologases I (r. ca. 51-78 CE) as part of the urban conglomeration of Seleucia-Ctesiphon (now known as Al-Mada'in). 451 The other three cities may be those listed by Prolemy: Apamea near the Tigris in Mesopotamia,“μετὰ τὴν Σελεύκειαν” (near Seleucia); Teredon in Babylonia, also near the Tigris; and Apollonia in the inland areas of Assyria. 452 This short list of Assyrian cities is followed by a more expansive list of the three most illustrious cities in the region, Babylon, Ctesiphon and Seleucia, all which were unambiguously within Persian territory . 453 Seleucia on the Tigris and Ctesiphon were regularly the target of Roman raids down the Tigris or Euphrates, but even when the Roman conquerors attempted to hold them, their possession was never lasting. As well as the Tigris and other rivers mentioned in his description of Adiabene, Ammianus lists three rivers in Assyria: the Marses, the flumen regium and the Euphrates; the first two names probably refer to the same river, the Naarmalcha canal. 454 Ammianus’ description of Assyria establishes the general location of the region by describing its contents, and establishes its topological relationship to the adjacent provinces and regions, but does not topographically locate its boundaries. This can be established to a certain degree by Ammianus’ account of cognominata et Teredon et Apollonia et V ologessia hisque similes multae.” (However, in all Assyria there are many cities, among which Apamea, named Mesene, Teredon, Apollonia and Vologessia and many others stand out.) 451 Kröger, Jens, ‘Ctesiphon’, EIran. Seleucia on the Tigris: M.-L. Chaumont (1984) 101–104. 452 Apamea: Ptol. Geog. 5.18.9. Teredon: Ptol. Geog. 5.20.5. Apollonia: Ptol. Geog. 6.1.6. 453 Amm. Marc. 23.6.23: “splendidissimae vero et pervulgatae hae solae sunt tres: Babylon cuius moenia bitumine Samiramis struxit — arcem enim antiquissimus rex condidit Belus — et Ctesiphon quam Vardanes temporibus priscis instituit, posteaque rex Pacorus incolarum viribus amplificatam et moenibus Graeco indito nomine, Persidis effecit specimen summum. post hanc Seleucia ambitiosum opus Nicatoris Seleuci.” (But these three alone are the most illustrious and populous: Babylon, whose walls Semiramis built with bitumen (for the ancient king Belus built the citadel), and Ctesiphon, which Vardanes founded long ago; and later king Pacorus strengthened it with additional inhabitants and with walls, gave it a Greek name, and made it the crowning glory of Persia. And after this is Seleucia, the ostentatious work of Seleucus Nicator.) 454 Amm. Marc. 23.6.25: “Perfluunt autem has easdem terras potiores ante alios amnes hi, quos praediximus, et Marses et Flumen regium et Euphrates cunctis excellens. qui tripertitus navigabilis per omnes est rivos, insulasque circumfluens, et arva cultorum industria diligentius rigans vomeri et gignendis arbustis habilia facit.” (But more important that others are these rivers already mentioned flowing through these same lands: the Marses, the Royal River and, greatest of all, the Euphrates, which divides in three, all entirely navigable, flows around islands and attentively watering the fields through the diligence of the farmers makes them ready for the plough and for bearing fruit trees.) For the Naarmalcha canal, see Appendix 1.12. 135/448 Julian’s expedition in 363 CE within which the description of Persian regions is situated. In fact, the location of this excursus is one indication of part of the boundary between Roman and Persian space. The Geography of Julian’s campaign (363-364 CE) Ammianus’ account of Julian’s campaign traces the entire route in some geographical and historical detail from the army’s departure from Antioch under Julian in 363 CE to its return to Nisibis under Jovian in 364 CE. Ammianus begins to give geographical details when the route moves between Hierapolis in Euphratensis and Batnae in Osrhoene (or as Ammianus calls it, Osdroene). Accordingly , it gives some sense of the Roman provinces across the Euphrates as well as showing the boundary between Roman and Persian space. Beyond Heirapolis, Ammianus refers to a number of locations in Roman space: Batnae, Carrhae, Davnae, the Balikh river, Callinicum, Cercesium, and the Khabur River. After Julian crosses into “Assyria”, Ammianus informs us about a second set of locations along the Euphrates and in Babylonia: Zaitha, Dura Europus, Anatha, Thilutha, Achaiachala, Baraxmalcha, Diacira, Macepracta, the Naarmalcha canal, Pirisabora, Maozamalcha, Coche/Seleucia, and Ctesiphon. The narrative of the retreat from Ctesiphon includes a final set of locations: the districts of Hucumbra and Maranga, then, after Julian’s death, the toponyms Sumere, Charcham, Dura on the Tigris, Hatra, Ur, Thilsaphata and Nisibis. 455 The boundary between Roman and Persian space on the Euphrates is clearly marked. When the narrative reaches Circesium, Ammianus notes: [Imperator Cercusium] ingressus est munimentum tutissimum et fabre politum. cuius moenia Abora et Euphrates ambiunt flumina velut spatium insulare fingentes. [2] quod Diocletianus exiguum antehac et suspectum muris turribusque circumdedit celsis, cum in ipsis barbarorum confiniis interiores limites ordinaret, ne vagarentur per Syriam Persae ita ut paucis ante annis cum magnis provinciarum contigerat damnis. 456 455 Some of these sites are described in Appendix 1. 456 Amm. Marc. 23.5.1-2: “The Emperor marched to Circesium, a very secure and skilfully built fortification. The Khabur and Euphrates rivers flow around the walls making a sort of island. [2] The place was small and of dubious strength before Diocletian surrounded it with high walls and towers, when he was arranging the interior of the frontier regions at the very edges of the Barbarian territory , so that the Persians might not wander through 136/448 This strong Diocletianic fortification (munimentum tutissimum et fabre politum) lay at the confluence of the Khabur and the Euphrates and marked the edge of Roman space. 457 Ammianus clearly marks the Khabur as the boundary of Roman space in his narrative. After crossing the Khabur, Julian address his troops, then sets foot into Assyria at dawn the next morning (“candente iam luce Assyrios fines ingressus”). 458 Between the speech and the dramatic start to the invasion proper, Ammianus inserts his lengthy survey of the Persian Syria as had happened a few years earlier with great damage to the provinces.” 457 See Appendix 1.8. 458 Julian’s speech: Amm. Marc. 23.5.16-25. Entry to Assyria: 24.1.1. 137/448 Map 21: Known Sites from Julian's Campaign empire, discussed above. Sundwall claims that this digression “ill-fits the immediate context of the chapter”; 459 on the contrary , the digression fits the narrative perfectly , and signals the crossing from Roman to Persian space. Such a lengthy digression was necessary for Ammianus to clearly signal this boundary because he had already begun to describe Julian’s journey into Assyria earlier in book 23. Ammianus’ work in general is filled with references to omens, magical practices and religious rituals. He is particularly interested in predictions of imperial deaths and his account of Julian’s march into Assyria abounds with ill omens. When Julian reaches Circesium, Ammianus describes an omen there, then immediately begins to describe the next part of the journey to Zaitha where Julian sacrifices at the site of Gordian’s death, then to Dura Europus where Julian receives another omen in the form of a dead lion, then several more omens as the journey continues. 460 However, Zaitha and Dura Europus lie south of the Khabur in Assyria. After the last omen in this set, Ammianus’ narrative returns to the bridge at Circesium, has Julian give his speech, gives his Persian geographical survey , and has Julian officially enter Assyria at dawn. In this way , Ammianus clearly marks the boundary between Roman and Sasanian space before proceeding to repeat the army’s arrival at Dura Europus at the start of book 24, this time in the correct place. 461 The second group of locations are those which Ammianus locates in Assyria. Most of these are clearly marked as belonging to the Sasanid sphere either directly by Ammianus’ text, or indirectly by the actions of the Roman army towards them. Ammianus describes Zaitha simply as a place (locus); its location is unknown. 462 The only feature Ammianus records at the site is the tomb of Gordian (Gordiani imperatoris... tumulum); perhaps indicating that by Ammianus’ time the site was only a village, if it remained 459 Sundwall (1996) 624. 460 Amm. Marc. 23.5.4-14. 461 Amm. Marc. 24.1.5; Matthews (1989) 130–31, 178–79. 462 Amm. Marc. 23.5.7. Gawlikowski (2007) 132–34. Ptolemy (Geog. 5.18.6) lists “Ζείθα (ἢ Ζήθα)” as one of the “cities or villages of Mesopotamia near the Euphrates” (“Πόλεις δέ εἰσιν ἐν τῇ Μεσοποταμίᾳ καὶ κῶμαι, παρὰ μὲν τὸν Εὐφράτην ποταμὸν αἵδε...”). 138/448 occupied at all. 463 Dura Europus appears as a deserted site, either a town or city (Duram desertum oppidum and civitatem... Duram desertam), on the bank of the Euphrates (marginibus amnis inpositam). 464 The first inhabited site encountered after crossing the Khabur is Anatha. 465 Julian sends a force to capture it, but even after the inhabitants agree to terms, the Romans burn the town. 466 This is clearly non-Roman territory . Thereafter, Ammianus describes a series of hostile encounters with fortresses and cities along the route (Thilutha, Achaiachala, Baraxmalcha, Diacira) 467 until they reach the heart of Sasanian Babylonia in the 463 Musil (1927) 337–39; M.-L. Chaumont (1984) 89. 464 Amm. Marc. 23.5.8; 24.1.5. See Appendix 1.9. 465 Anatha was an island fortress, four days from Dura; Amm. Marc. 24.1.6: “Anathan munimentum... quod ut pleraque alia circumluitur fluentis Euphratis” (the fortification of Anatha... which, like many others, is surrounded by the stream of the Euphrates.) Musil (1927) 345–49; Gawlikowski (1988) 90; Ibrahim (1986) 80. Isidore calls the same place Anatho and places it twenty-seven schoinoi along the route. Anatha (also known as Ana) was held by Rome in the third century CE, and probably by the Palmyrenes as early as the first. Since 1985, the ancient site now lies beneath the waters of the northern end of the Haditha reservoir in Iraq near the relocated modern town, Anah. 466 Amm. Marc. 24.1.6-9. Musil interprets the surrender of the fort, the rewarding of the commander and the deportation of the population to Calchis as evidence of collusion between the inhabitants and the Romans, Musil (1927) 238. However, that seems to ignore the possibility of forced relocation. 467 Thilutha is Telbis Island: Amm. Marc. 24.2.1: “...castra pervenimus nomine Thilutha in medio fluminis sita, locum inmenso quodam vertice tumescentem et potestate naturae velut manu circumsaeptum humana...” (W e came to a fortress by the name of Thilutha situated in the middle of the river, a place rising up to a high summit and surrounded with the power of nature as if by a human hand...); Isidore 1: “μεθ᾽ ἣν Θιλαβοὺς νῆσος κατὰ τὸν Εὐφράτην, ἔνθα γάζα Πάρθων, σχοῖνοι β\.” (After which is Thilabous, an island in the Euphrates, where there is a treasury of the Parthians, two schoinoi.) M.-L. Chaumont (1984) 97; Gawlikowski (1988) 85, 90. On Telbis island, now inundated by the waters of the Haditha Dam: Musil (1927) 239. A brief description of the rescue excavations can be found in Killick and Roaf (1983) 222. Of note is evidence of Parthian occupation, but “as yet no sign of the royal treasure mentioned by Isidore of Charax”. Captain Chesney’s Euphrates expedition of 1831 reported seeing “the still perfect walls of ancient Tilbus” on the island, quoted in Matthews (1989) 146. Guest (1992) offers an engaging account of Chesney’s ill-fated attempt to prove the utility of the Euphrates for delivery of mail between the UK and British India. Achaiachala: Amm. Marc. 24.2.2: “ quo transito cum ad munimentum aliud Achaiachalam nomine venissemus fluminis circumitione vallatum arduumque transcensu, refutati pari responso discessimus. alia postridie castra ob muros invalidos derelicta praetereuntur incensa.” (After passing that place, when we came to a another fortress named Achaiachala, fortified by the surrounding river and difficult to ascend, rebuffed in the same way , we departed. The next day a fortress abandoned because of its crumbling walls was burned in passing.) Musil indentifies Achaiachala with the town of al-Hadita, Musil (1927) 239. Baraxmalcha is probably to be indentified with Bidjan Island and Isidore’s Izan (῎Ιζαν): Amm. Marc. 24.2.3. Isidore 1: “εἶτα ῎Ιζαν νησόπολις, σχοῖνοι ιβ. εἶτα ᾽Αείπολις, ἔνθα ἀσφαλτίτιδες πηγαί, σχοῖνοι ις.” (Then Izan, and island city , 12 schoinoi, then Aipolis, where there are springs of asphalt, 16 schoinoi.) Bijan Island is another site flooded by the Haditha Dam. It was occupied by Parthians, (perhaps Palmyrenes), and Romans sequentially until the mid-third century CE. Edwell (2008) 73–74; Killick (1983) 208; M.-L. Chaumont (1984) 97–98; Gawlikowski (1988) 85; Ibrahim (1986) 79. Diacira is Hīt; the Assyrian name for the place, Id, is derived from the word for pitch, and the Aramaic name, Ihi Dikira, means ‘place of asphalt’: Matthews (1989) 147–48; Musil (1927) 239, 350–53. Isidore gives Diacira/Hīt the name Aipolis, notes the bitumen springs, and places it sixteen schoinoi beyond Izan: M.-L. Chaumont (1984) 98–99; Gawlikowski (1988) 85, 90. This part of 139/448 vicinity of the city of Macepracta near the inflow of the Naarmalcha canal. 468 The return of the Roman army by a different route, towards Nisibis, allows us to investigate the border between Roman Mesopotamia and Sasanian Assyria in another location. After breaking off the invasion, the Roman force proceeds up the Tigris towards Corduene (Gordyene). 469 The army passed the city of Dura on the Tigris (“civitatem nomine Duram”) before crossing that river and reaching the abandoned ruins of Hatra: 470 ...prope Hatram venimus, vetus oppidum in media solitudine positum, olimque desertum, quod eruendum adorti temporibus variis Traianus et Severus principes bellicosi cum exercitibus paene deleti sunt, ut in eorum actibus has quoque digessimus partes. 471 This is one of the few literary accounts of Hatra which, like Dura Europus, had been deserted for over a century before Ammianus’ arrival. 472 After at least six days and seventy miles, Jovian’s army reached a Persian fortress called Ur (“Ur nomine Persicum... castellum”). 473 Dillemann suggests this may have been Zagura, perhaps Orurus, the border of the empire in Pompey’s day , according to Pliny . 474 Ammianus describes Ur as a Persian fortress, which he probably meant in historical rather than political terms. In other words, it was Persicum in that it had been built by the Persians, not Persicum in that the Persians currently controlled it. Babylonia was renowned for its bitumen springs; both Strabo (16.1.15) and Ammianus (23.6.16) describe the production and qualities of naphtha in southern Mesopotamia. 468 For the canal and the city , see Appendix 1.12. 469 Ammianus attributes the choice of route to the will of the gods, as established by traditional Roman sacrificial practice: Amm. Marc. 24.8.4-5. 470 The army passes a village (villam) named Hucumbra (Amm. Marc. 25.1.4) an area (tractum) named Maranga (Amm. Marc. 25.1.10-11), then, after Julian’s death (Amm. Marc. 25.3) a fortress (castellum) named Sumere (Amm. Marc. 25.6.4) a place (locus) called Charcha (Amm. Marc. 25.6.8), then Dura on the Tigris (Amm. Marc. 25.6.9). 471 Amm. Marc. 25.8.5: “W e arrived near Hatra, an old town situated in the middle of a desert, and long since abandoned. The warlike emperors Trajan and Severus attempted to destroy it with their armies on various occasions and were almost destroyed themselves, as I described in my accounts of their reigns.” Trajan’s attempt is described by Dio Cassius (68.31). 472 Hauser, Stefan R., ‘Hatra’ BNP; ‘Hatra’ PECS, pp.379-80; Dirven (2013); Drijvers (1978). In particular, Dirven’s edited volume describes the history and current state of research on the site. Isaac’s contribution (2013) collects the literary sources for Hatra. 473 Amm. Marc. 25.8.6-7. 474 Dillemann (1962) 311. Pliny , NH 6.120. 140/448 After all, Jovian’s retreating army was met at Ur by a Roman tribune sent with supplies from the Roman forces in Mesopotamia. There is no mention of a sizeable military force which might be expected to accompany a supply train to a meeting in the shadow of an enemy fort. Although Ammianus does not relate any conflicts in the desert, the Roman army had been harried by Persian forces even after the peace treaty signed on the Tigris, but there is no reference to fighting, hostilities or even wariness at Ur. The peaceful encounter there with the supply train probably indicates that this was formerly a Persian fortress which at this time lay under Roman control; it may have been one of the fifteen forts that Jovian had agreed to hand over to the Persians in his treaty . 475 Closer to Nisibis, Ammianus mentions a place called Thilsaphata, but without clarifying what kind of place it is. 476 Dillemann identifies Thilsaphata with Pliny’s Thebata, but neither site has been identified archaeologically . 477 The final acts of the campaign take place at Nisibis, but Ammianus gives no geographical information about the city . 478 Nisibis was certainly part of the Roman province of Mesopotamia at this time and the Roman forces charged with defending Mesopotamia met the retreating army at Thilsaphata, so it too was probably within their jurisdiction. The desert region between the Tigris and Ur probably formed the boundary between Roman and Persian territory in this area. Ammianus gives no indication that there was any development of this space that would have required a more precise division of space. Conclusion These three geographical sections within Ammianus’ historical work show Ammianus’ approach to geography and to the Mesopotamian borderland. As a geographical writer, Ammianus is mostly concerned to contextualise the historical events he describes. Where he embarks on significant geographical digressions, it is to describe a place or people relevant to an impending conflict. That his geography is in service of 475 Amm. Marc. 25.7.9. 476 Amm. Marc. 25.8.16. 477 Dillemann (1962) 311–12. See above, p.56. 478 Amm. Marc. 25.8.17; 25.9.8. 141/448 history makes it no less valuable. Ammianus was a keen observer of events, situations and space and provides valuable geographic information about the Mesopotamian borderland. His focus lay on the places in which history occurred, especially cities and sites important to his narrative goals. Into the latter category fell places like the crossing-point of the Khabur, ripe with liminal symbolism, and the sites where Julian witnessed omens of his death. Moreover, Ammianus’ geographical descriptions show the relatively strict division of space between the two empires which had developed over the subsequent four centuries, a topic to which I will return in Chapters Two and Four. Defining and Describing the Mesopotamian Borderland I will conclude this chapter by bringing to bear two theoretical frameworks from the introduction. First, recalling Bradley Parker’s borderland framework, I will discuss the geographical boundaries evident in the descriptions of “Mesopotamia” as a geographical area, that is, northern Mesopotamia. The defining characteristics of a borderland are that it is a zone and that it contains multiple overlapping sets of boundaries. 479 This chapter has been primarily concerned with the representation of geographical boundaries, but while all of the works discussed above include geographical descriptions of the Mesopotamian borderland, they do not all take a perspective on boundaries and the delimitation and denomination of space that we might consider primarily geographical in the sense of Parker’s typology . Following that, I will examine the deployment of the two geographic modes of linear and planar geography . In the introduction, I noted that I would examine the existence of these modes in the narrative structures of the borderland. I will conclude this chapter by drawing together the threads of that examination which run through this chapter. Strabo’s approach is primarily geographical. He names Mesopotamia from its position between the Euphrates and Tigris and delimits the region according to those rivers, the Taurus, and the desert which separates Mesopotamia and Babylonia. Within that space, he names areas according to ethno-cultural 479 See Introduction. 142/448 markers (Gordyene and the Mygdonians) and locates them according to their topological relationship to geographical features (Gordyene in the mountains, Mygdonia near the mountains and the river, the Skenitai in the desert). At a lower level still, he often locates specific sites according to their topological relationship with nearby features (Nisibis under Mount Masius, Anthemusia, Bambyke and Skenai near rivers or canals). Pliny’s narrative takes the geography of the region into consideration, but is primarily structured according to broader political concerns. 480 Like Strabo, he names the region for the rivers and uses those as the major delimitations for the space. Pliny does not specifically make the Taurus the northern boundary of Mesopotamia, but the places and people he locates within the region imply that his conception of its northern limit also falls along that range. To the south, Pliny is particularly vague. Pliny’s approach to internal divisions differs between the two books in which he treats the region. In book five, he names spaces using a mix of ethnic (the Orroeni and Praetavi Arabs) and administrative (praefectura Mesopotamiae) terms, and provides no clear delimitation of their respective boundaries beyond those implied by the definition of Mesopotamia itself. In book six, he organises the spatial sub-division of Mesopotamia by lists with reasonably clear internally topography , but which are not topologically linked to each other. Those lists are often denominated along ethnic lines (such as the list of Arabian tribes at 6.117-18 or the list of mountain tribes at 6.118). Ptolemy’s account is also primarily geographical, but in a different way to that of Strabo. His work does not elaborate on the etymological details of the places he describes, but the boundaries he assigns to Mesopotamia imply that the standard denomination according to the two rivers applies here also. Those rivers form the usual boundaries while he specifies the southern boundary according to a third, the Naarmalcha canal which links the Euphrates to the Tigris in northern Babylonia. In the north, the Taurus is the boundary between Mesopotamia and Armenia as usual. Within Mesopotamia, internal districts are 480 For the way these political concerns are reflected in Pliny’s representation of Mesopotamia, see Chapter 4. 143/448 topologically linked to each other, and to the surrounding geography (Anthemusia is near Armenia, Akabene is near the Tigris and Ankobaritis near the Euphrates). Only some of the cities of Mesopotamia are related to a geographical feature (namely those noted as near the Euphrates and Tigris rivers), but all are topologically located to each other through coordinate geometry . The Expositio Totius Mundi names the region using administrative categories (Mesopotamia and Osdroena), but treats them as a group implicitly governed by the geographical boundaries of the two rivers. It neither describes the physical geography of Mesopotamia, nor topologically relates the places and spaces it mentions to each other or the wider space. The Expositio’s description includes the people who inhabit those two provinces, but in an economic rather than an ethnographic or cultural way . Ammianus includes many geographical details in his work, but his descriptions of spaces at a regional scale are largely based on topological relationships to other administrative units. The Euphrates acts as a border between Syria and Osdroena, but the Tigris is internal to the Persian administrative region of Assyria. He places the border between Mesopotamia and Assyria along the Khabur at least at Circesium, but his prioritising of political and administrative boundaries means the geography of the space is only as useful to Ammianus as it was to the Roman administrators who assigned those boundaries. While the Euphrates and Tigris rivers are most commonly used to define the geographical space of Mesopotamia, the northern and southern edges are more nebulous. To the north, the boundary was marked consistently by the Taurus. However, while a river comprises a relatively consistent and narrowly defined line, mountains are much more difficult. In his account of the history of the Pyrenees as a boundary between France and Spain in the early modern period, Peter Sahlins shows how the existence and status of mountain valleys complicates state attempts to define precise boundaries of control in mountainous areas. 481 This problem has already been seen in relation to Strabo’s placement of Tigranocerta. Tigranes’ capital lay to the 481 Sahlins (1989). 144/448 north of the Tigris in the area of the Transtrigratine provinces of Armenia which were exchanged between Rome and Sasanid Persia in the treaties of 298 and 363 CE. 482 Syme considered the northern plateau between the Tur Abdin/Masius and the Taurus range proper (that is, the upper Tigris basin) to be part of Armenia, but Strabo considered it part of Mesopotamia. 483 For Strabo, it is that latter range that divides Mesopotamia from Armenia, not the Tigris. The upper Tigris valley , and the Gordyaean mountains (including Tur Abdin/Masius) comprise Strabo’s Gordyene. Strabo is not bound by the strict divisions implied by the etymology of the word “Mesopotamia”. In his account of the political allegiances of the Mesopotamian Skenitai, he discusses peoples both north and south of the Euphrates. 484 Strabo’s Mesopotamia includes both the southern slopes of the Taurus north of the Tigris and steppe and desert regions on the south side of the Euphrates. Pliny is not specific about the northern boundary , but in book six he shows a similar approach to Strabo when he describes mountain tribes in Gordyene and Adiabene along with the tribes of the fertile and desert regions to the south. Ptolemy attempts to define Mesopotamia with the greatest precision; the boundaries between his regions always include coordinate locations. Accordingly he gives precise coordinates for the places where the Euphrates and Tigris rivers intersect that imaginary line of the Taurus on their journey downstream from Armenia sources. His conception of Mesopotamia’s borders is the strictest. The southern boundary , that between Mesopotamia and Babylonia is even more unclear. Ptolemy is again precise, giving coordinates locating the Naarmalcha canal as the boundary . However, this is further south than previous writers. Ptolemy is alone in locating Seleucia on the Tigris within Mesopotamia rather than in Babylonia where is it usually placed. Ptolemy’s desire for geographical precision led him to use bodies of water to divide space whenever possible, at the expense of other methods of aggregating space. The more common distinction between Mesopotamia and Babylonia probably stems from farming method, 482 For which, see Chapter 2. 483 See above, p.43. 484 Strabo 16.1.28. Chapter 5. 145/448 although this is not noted explicitly . The northern regions (Mesopotamia) are characterised by dryland farming in which the fields are watered primarily by rain, while in the south (Babylonia), an ancient system of canal-based irrigation provides the bulk of the agricultural water. Seleucia falls within the irrigated zone and (along with Ctesiphon across the Tigris) is usually described as the most important city in Parthian Babylonia. 485 Strabo does not give a precise division, although he does conform to the agricultural distinction just proposed. He includes the steppe pastoralists and communities along the Euphrates strip in Mesopotamia and the irrigated lands to the south (including Seleucia and Ctesiphon) in Babylonia. Pliny is even less precise. Although he shows more knowledge of the various tribes that inhabit Mesopotamia (broadly defined to include mountain tribes of the Taurus and Zagros and perhaps Skenitai near Mesene), his descriptions of Mesopotamia, Adiabene, Babylonia and the regions to the south blend seamlessly together without delimitation. He knows of and names the Naarmalcha canal, but does not attribute it any conceptual status as a boundary as Ptolemy did. The geographical space between the Euphrates and Tigris was certainly considered a distinct geophysical region by the geographical writers of the Roman empire and distinguished from the more intensively developed southern region of Babylonia. In those cases (Pliny and Ammianus) where the distinction between the geographical region which the Roman writers called Mesopotamia and that which they called Babylonia was relatively weak, political concerns are clear in their presentation. In the case of Ammianus, his main method of denominating and delimiting space was according to political and administrative boundaries and categories (the Roman provinces of Osdroena and Mesopotamia and the Persian region of Assyria). Pliny’s political treatment of the Mesopotamian borderland is more complex and will be discussed in chapter four. As noted in the introduction, it has become commonplace to describe the Roman conception of 485 Ptolemy places Ctesiphon (on the east side of the Tigris) in Assyria rather than Babylonia, Geog. 6.1.3. 146/448 space and movement as linear and hodological and some consideration of this aspect of their narrative treatment of Mesopotamia is warranted. This is an issue of narrative and descriptive topology; how places are located relative to each other and how they are described as such in the narrative. Although many texts and narrative segments display a linear understanding of space, the imperial geographic writers frequently structured and delimited space using planar, “map-like”, conceptions of the world. Such a view dominates the scientific geography of Claudius Ptolemy (as we might expect for a text explicitly described as instructions for making a map), but it is evident in the less technical geographical works as well and serves as a reminder that there were many ways of understanding geographic space in the ancient world. Each of these five authors establishes the topological relationships between the places they discuss in a quite different way , each appropriate to their geographical project. Strabo’s approach is consistent throughout his work. He defines the regional area to be described by relation to geographical features, often divides the broad area into smaller spaces also by geographical features, then works his way systematically through the smaller areas until the entire region has been described. In the Mesopotamian borderland, he describes Assyria as a whole, then the sub-regions which comprise it, including Mesopotamia. He divides Mesopotamia into three parts using relative topology , physical geography and ethnography as references: One part is along the mountains (ἡ παρόρειος), another is inhabited by the Gordyeans (τὰ τῶν Γ ορδυαίων χωρία), and the last is towards the south and further from the mountains (τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα καὶ ἀπωτέρω τῶν ὀρῶν). 486 This descriptive approach is rooted in the two-dimensional topology of the space. 487 Parts of the area are near the Euphrates (πρὸς τῷ Εὐφράτῃ), Nisibis is at the foot of mount 486 Strabo 16.1.23, 24, 26. 487 Strabo often structures his descriptions along lines of communications such as those afforded by rivers, roads or coastlines, Dueck (2000) 167–8. His description of Mesopotamia does not use either river in this way; rather, it has a clear sense of two-dimensionality. His work generally mentions topological relationships between the spaces he describes, often in the initial survey. For example, see his initial survey of οἱ Ἀσσύριοι (16.1) which is bordered by (συνάπτουσιν) Persis and Susiana and stretches as far as the Cilicians, Phoenecians and Judaeans (μέχρι Κιλίκων καὶ Φοινίκων καὶ Ἰουδαίων). 147/448 Masius (ὑπὸ τῷ Μασίῳ ὄρει), and the river Aborras flows near Anthemousia (περὶ τὴν Ἀνθεμουσίαν). 488 By relating features and places to each other in this way , Strabo allows the reader to imagine themselves moving and viewing within the space. By emphasising the relationships between geographical features, this topological approach constructs a physical and political geography based on spatial connections and movement. Strabo's Mesopotamia is a geographical space interrelated at the local, regional, and supra- regional levels. 489 Pliny provides topological information in a quite different way to Strabo. Although Strabo organises parts of his work as itineraries or periploi, there are moments of two dimensional, planar, description throughout. By contrast, Pliny’s geographical work is fundamentally linear. This is most obvious in his description of Syria, which consists of an linear narrative up the coast and down the interior, broken only by the two strictly atopological lists of north Syrian cities and peoples. The part of Pliny’s Mesopotamian narrative included with his description of Syria in book five is based around a list of sub-regions (Osrhoene/Orroeni Arabia, praefectura Mesopotamiae and the land of the Praetavi Arabs) with atopologically listed contents (Edessa and Carrhae, Nikephorion and Anthemusia, and Singara). The first item of this list (Osrhoene) is located in space by a topological relationship (it faces Commagene across the Euphrates), while the rest simply appear as a list which follow Osrhoene. In book six, Mesopotamia is also structured as a series of lists. Whereas the two north Syrian lists are interspersed with descriptive asides which act to hide their alphabetical nature, the series of lists which comprise Pliny’s description of Mesopotamia is interspersed with spatial dislocations: the list of Arab tribes ends somewhere in Babylonia, then a new list of tribes on the fringes of Taurus begins in the foothills of the Tigris. None of these lists are identified; the “list of Arab tribes” is not named as such, nor is the list of mountain tribes. The tribes that comprise these lists are linked to each other and occasionally to independent topographical features, but not with sufficient detail 488 Strabo 16.1.23, 27. 489 At the supra-regional level, see Strabo’s construction of Assyria, Strabo 16.1.1. 148/448 that they could be located in space if not for those topological connections to the other tribes on the list. Ptolemy’s Geography is highly topological, but the mathematical basis of his work means that his account is of a different nature to the other Imperial geographic writers. At the broadest level, each region is topologically located by relationships of adjacency based on a combination of cardinal directions and geographical features; Mesopotamia is described as adjacent to Syria, Armenia, Assyria, Babylonia and Arabia Deserta. 490 Similar relationships are often established between districts within each region. 491 While this is the case in Mesopotamia, where Anthemusia, Chalkitis, Gauzanitis, Axabene, [T]ingene, and Ankobaritis are all topologically interrelated, the districts in Syria are related to the features within them rather than to each other. The individual coordinate locations which make up the bulk of Ptolemy’s work also act to topologically relate the sites to each other, but at an order removed. That is, he provides the tools for the reader to locate those features but the relationships are not apparent until the reader recreates Ptolemy’s map itself. Much like Strabo, Ammianus uses topological relationships to establish a clear sense of two- dimensionality in his geographical narratives. 492 He begins his geographical descriptions by orienting the reader, such as in his description of the eastern provinces of the Roman Empire which begins with an imaginary crossing of the Taurus (“superatis Tauri montis verticibus... Cicilia spatiis porrigitur late distentis...”). 493 He then uses relative topological references to move through the space described by his narrative (in the same case, “...eiusque lateri dextro annexa Isauria...”). 494 The spatial organisation of each geographical passage varies according to its purpose in the overall narrative. In his descriptions of the Roman provinces and Persian regiones, topological references convey a good sense of the two-dimensional space, 490 Ptol. Geog. 5.18.1.Yes 491 Ptol. Geog. 5.18.4. 492 See the excellent analysis of Sundwall (1996) esp. 633ff. 493 Amm. Marc. 14.8.1: “When the summits of the Taurus have been crossed... Cilicia extends widely in an open plain...” Sundwall discusses this passage at (1996) 633–34. 494 Amm. Marc. 14.8.1: “...and Isauria, connected on the right side of [Cilicia]...” 149/448 while in his account of Julian’s march, he appropriately links geographical space in a linear fashion along the route. The author of the Expositio understands two-dimensional spatial structure, but is generally unconcerned with it. His understanding can be shown by his concern to demonstrate the westerly direction (“ad occidentem”) of his initial description of the peoples beyond Roman space, 495 and later, between his descriptions of Syria and Egypt, he indicates that Egypt lay to the left of Syria (“de laeva parte Syriae”). 496 This reveals the familiar conception of geographical orientation which places north at the top of a map, with west as left and east as right. This conception is also seen after the Expositio’s lengthy and less-disjointed description of Egypt. 497 When the narrative moves east, Arabia is described as “de dextris... Syriae”. 498 However, as in his description of Mesopotamia, the author provides no internal topological structure to the provinces he describes. These five accounts of Mesopotamian space show the variety of geographical methods that were available to the Roman Imperial geographic writers. This variety included not only the specifics of exactly where the boundaries of the chosen space were placed, but whether they were defined in geographical, political or ethno-cultural terms. Judging by the frequency with which they appear and the detail with which they were described, cities were the most important kind of place in ancient Roman geography . The case of Nisibis illustrates the variety of topological approaches to locating cities among these authors. In Strabo, Nisibis is located within an internal district of Mesopotamia (Mygdonia) and topologically linked to a nearby geographical feature (Masius), in the Expositio it floats adrift in the provincia Mesopotamia, while in Ptolemy , it is somewhat adrift in the middle region between the rivers, but nevertheless connected to the rest 495 Expositio 8. 496 Expositio, 34: “Habes ergo de laeva parte Syriae et Aegypti et Alexandriae et totius Thebaidis partes, quas describere necessarium est.” (Therefore on the left of Syria you have Egypt, Alexandria and the whole Thebaid; these must be described.) 497 Expositio, 34-37. Alexandria is a special focus of this passage. 498 Expositio, 38. 150/448 of Mesopotamia by its common attachment to Ptolemy’s coordinate plane. 499 As in Strabo, the same city can be treated in quite different ways in the same work. In book six of Pliny’s Natural History , Nisibis appears twice, once in a topological relationship to a tribe in a linear list of tribes and once as the unattached capital of Mygdonia (itself curiously aggregated to Adiabene). 500 In the extant books of Ammianus it appears a number of times in a variety of contexts and topologies, where it is related to the narrative as appropriate, but without much geographical information. 501 The case of Nisibis is indicative of the varied treatment that the Mesopotamian borderland and the places and spaces contained within it receive in the geographical sources. 502 Moreover, it is indicative of the way that the geographical writers might not select certain sites, parts of the region, or aspects of the borderland as worthy of representation in their construction of the space. 499 Strabo 16.1.23; Expositio 22; Ptol. Geog. 5.18.11. 500 Pliny NH. 6.117 (in a list of Arab tribes); 6.42 (as the capital of Mygdonia). Nisibis does not appear in book five, despite the coverage of that book extending as far east as Singara, southeast of Nisibis. 501 Nisibis appears 16 times in 11 passages: Amm. Marc. 14.9; 18.6; 18.7; 19.9; 20.6; 20.7; 23.6; 25.1; 25.7; 25.8; 25.9. 502 The political and economic roles of Nisibis in the borderland will be examined in Chapters 2 and 5. 151/448 Chapter 2: States in the Borderland In Chapter 1, I examined the denomination and delimitation of the physical space of the Mesopotamian borderland in the Roman imperial geographic writers. In this chapter, I zoom in to analyse the local political and social structures that those authors describe in their work. I am concerned on one hand with the denomination and spatial delimitation of these political entities, and on the other with the representation of dynamic processes which link them to each other and the Roman and Iranian worlds, in particular the representation of political networks in the descriptions of Commagene, Osrhoene, Nisibis and Palmyra. The Roman imperial geographic writers often treat political structures in ambiguous terms. Even when the narrative portrays these structures as firmly fixed in terms of their political allegiance, details in the texts undercut and problematise those allegiances. Because these network connections participate in dynamic processes, a sense of historical time is required. Accordingly , I will outline the historical events and processes which precede the geographical representations of these kingdoms and cities in order to cast light on those representations. I will begin with a discussion of the network theory of empires: the control of space by localised points of power rather than by the broad control of territorial space. Network Empires The Problem There is recognition among geographers and cartographers that maps with clearly defined border lines give a misleading impression of territorial control. 503 Cartographic representations lend themselves to the assignment of apparently linear divisions of space more than narrative geographic texts, but borders and boundaries are fluid and contingent. Pre-modern cartographic representations in particular may give 503 See the extensive bibliography in Smith (2005) 832–34. 152/448 spurious authority to the lines and boundaries whose reality they assert. This was often intentional. State ideological texts usually show the state they represent as a stable construct in which state power is evenly and ubiquitously distributed. 504 Geographical texts often participate in this construction. 505 The idea that lines could be drawn around Roman territory was certainly current in the Augustan age, even though such lines would not have been regarded as a limit to expansion. In the Res Gestae, Augustus boasts to have increased the fines (boundaries) of all the provinces. 506 Strabo frequently refers to the ὅρια (boundaries) of cities, provinces and empires in his work, including his notification of the border of the Parthian empire (τῆς Παρθυαίων ἀρχῆς) along the Euphrates. Strabo usually locates these boundaries in specific places or on geographical features (such as the Euphrates) rather than attempt a comprehensive description of the outlines of a space as we find in Ptolemy’s work. 507 Ideas of territory were often central to the self-conception of ancient states. The concept of territory is primarily an idealised vision of the real space over which the state’s control could be exerted. 508 Representations of space which remove heterogeneity guide our expectations of the social and administrative behaviour of the states which claim to control those spaces. 509 Chapter 4 will examine this at a broad scale. 504 Smith (2007) 32. 505 Nicolet (1991); Clarke (1999). 506 Res Gestae 26: “Omnium provinciarum populi Romani, quibus finitimae fuerunt gentes quae non parerent imperio nostro, fines auxi.” (I increased the boundaries of all the provinces of the Roman people which were adjacent to peoples not subject to our empire.) 507 Ptol. Geog. 15.18.1: “Ἡ Μεσοποταμία περιορίζεται ἀπὸ μὲν ἄρκτων τῷ ἐκτεθειμένῳ μέρει τῆς Μεγάλης Ἀρμενίας, ἀπὸ δὲ δύσεως τῷ ἐκτεθειμένῳ παρὰ τὴν Συρίαν τοῦ Εὐφράτου ποταμοῦ μέρει, ἀπὸ δὲ ἀνατολῶν τῷ παρὰ τὴν Ἀσσυρίαν μέρει τοῦ Τίγριδος ποταμοῦ τῷ ἀπὸ τοῦ πρὸς τῇ Ἀρμενίᾳ τμήματος μέχρι τῶν τοῦ Ἡρακλέους Βωμῶν... ἀπὸ δὲ μεσημβρίας τῷ λοιπῷ μέρει τοῦ Εὐφράτου παρὰ μὲν τὴν Ἔρημον Ἀραβίαν μέχρι πέρατος... παρὰ δὲ τὴν Βαβυλωνίαν μέχρι τῆς πρὸς τὸν Τίγριν συναφῆς καὶ τοὺς εἰρημένους Βωμοὺς...” (Mesopotamia is bounded on the north by the region of Greater Armenia already mentioned, on the west by the part of the Euphrates river near Syria (previously described), on the east by the part of the Tigris river near Assyria, from there to the part of Armenia as far as the altars of Heracles... To the south, the remaining part of the Euphrates near the Arabian Desert as far as navigable... And from Babylon up to the connection to the Tigris and the aforementioned altars...) 508 Smith (2005) 834–35. Steven Grosby (1995, 147) Territoriality: The transcendental, primordial feature of modern societies. Nations and Nationalism 1 (2): 143–62. Grosby 1997. Borders, territory and nationality in the ancient Near East and Armenia. Journal of the Economic and Social History of the Orient 40 (1): 1–29. 509 Smith (2007) 28; Konig and Whitmarsh (2007); Lewis and Wigen (1997). 153/448 Strabo assigned a border to the Parthian empire then implied a fuzzy edge to Roman territory . Pliny implied the extent of the two empires through his narrative structure but included a considerable area of overlapping space in both territories. Ptolemy avoided the question of imperial territory by focusing on geographic boundaries. The Expositio made a clear distinction in its narrative and thus implied a defined border on the ground, but that text was unconcerned with precise geographical representation of real space. Only in Ammianus do we find a relatively strict political border in a geographical text. Even there, it is only because one of the two places where it is crossed is spatially defined by a linear geographic feature (the Khabur river) that we can speak of a “strict border” in any sense. The other border section which is crossed (near Nisibis) appears as an arid zone rather than an identifiable line. Although at times the ambiguous representation of territory in these geographical narratives occurs as part of a deliberate ideological stance on the part of the geographic narrator, that narrative ambiguity also reflects a real territorial ambiguity . A states’ power on the ground was nebulous and difficult to encapsulate. 510 Just like geographical representations, state investment in territories was selective, strategic and heterogeneous. I do not mean “strategic” in the sense of an overarching grand strategy of empire, but in the sense that intentional decisions were made at various times to invest resources in specific areas of territory and specific boundaries. These decisions need not have been the subject of high level coordination by imperial administrators. States as Networks The notion of a territory across which state power is evenly present, available and applicable is a rhetorical fiction. Recent theorists of state and imperial control have begun to consider state power using a 510 I mean “power” in the most general sense defined by Mann (1986) 6: “power is the ability to pursue and attain goals through mastery of one’s environment.” I mean “state power” as that power (from whatever source) which acts in service of the state, whether deliberately wielded or as a emergent property of state actors; Mann’s concepts of extensive and intensive and authoritative and diffused power are important here, (1986) 7–10. 154/448 network model wherein power is disseminated from points in the landscape and is communicated and enforced along links between those nodes. 511 Anthropologist Monica Smith argues that a network model drawn from biological patterns of animal behaviours can improve our understanding of the internal dynamics of empires and states. 512 The core of the comparison rests on resource scarcity . Resources are not evenly distributed in nature. Animals spend most of their energy (an internal resource) claiming nodes of external resources and the connective corridors which allow movement between those nodes. They spend comparatively few resources maintaining territorial boundaries around that resource-movement network. More complex animals require larger networks spanning wider territories to satisfy all their resource requirements. The security of resources and resource exploitation within this network is a priority , not the security of territory as such. This is a flexible and energy efficient approach to landscape use with clear applications to the internal dynamics of states. Just like animals, states depend on resources for survival. Smith argues that people think in terms of resource nodes rather than homogeneous spaces. 513 States focus their internal resources at points where valued external resources are able to be collected, while areas of lesser value are invested with fewer state resources. In other words, a state’s power is focused on the spaces it values the most and spaces it does not value as highly are more open to alternative expressions of power, such as local autonomy or the power of a rival state. In considering complex human societies, our definition of resource can be expanded. A state’s internal resources can be administrative, economic, military , ideological, and so on, while the required external resources include human labour, food production, and the legitimation of existing power 511 Wilkinson (2002); Smith (2005) 838; Smith (2007). Sociologists have long recognised the power of networks as an explanatory model for various social phenomena, for example: Stark and Bainbridge (1985); Mann (1986) 1– 3. For a very approachable account of the history of network analysis, see Barabasi (2003). 512 Smith (2007) 29. 513 Smith (2007) 31. On movement between resource centres in a heterogeneous environment, see Johnson (1977) 488–94. In the context of pastoralist movement, see Turner (1999) and on Skenitai below , p.Error: Reference source not found. 155/448 structures. 514 The fewer internal resources are available to the state in the region in question, the more such a model applies. 515 Those resources will tend to be focused at certain sites or areas and absent elsewhere. Such a model is particularly appropriate in areas where the predominant terrain is of lesser value to the controlling state. Instruments of domination such as administrative and judicial centres, ideological resources, and military units are spread more thinly and in lower concentration so power is less able to be universally applied. This is not to say that state power only exists where state resources are present; the ability to project force and threaten intervention means that state power radiates from those points depending on the degree to which the inhabitants of more distant locations respect the ability and will of that state to exercise compulsion, or the degree to which they desire the benefits which the that state can offer. 516 The resources which ensure the state’s survival are more important than the maintenance of territorial boundaries. Boundary definition might still be an important consideration for states at particular times and in particular places, but a defended boundary is an external resource which is subject to the same selective considerations as any others within a state. In the context of the Roman empire of the Julio-Claudian period, the presence of the Parthian empire meant that boundary definition in the eastern provinces was given a high priority , with multiple visits by the imperial family and a comparatively high degree of militarisation. Compared to Mauritania, where such matters were mostly left to local Roman administrative officials to resolve with local 514 In fact, any of the sources of social power discussed by Mann (1986) apply. 515 Smith (2005) 844: “With limited resources to expend, central authorities assembling an initial state-level bureaucratic apparatus should have been particularly focused on the efficient use of nodes as places of investment that could be linked through cost-effective corridors of transport and communication.” Sinopoli (1994) 171–72. This idea underlies the “central place theory”: described in general by Johnson (1977) 494–501; applied to Seleucid Syria by Grainger (1990) 91–136. 516 Wilkinson et al. (2005) 24. Stein (1999) suggests that geographical distance is a major factor in interregional network variation (p.58). He criticises Wallerstein’s world-system model of core-periphery interaction (pp.10-43) and proposes a distance-parity model for interactions between a distant imperial core and the local regions it seeks to influence, control or dominate which puts the focus on the region itself rather than on the core. In this model, as distance to the core increases, parity between imperial and local interests increases. This model also applies at the intra-regional scale where nodes in a local network correspond to local “cores”, i.e. the influence of a given node decreases as distance increases (pp.175-76). This idea also underlies the experimentation with “gravity models” in archaeology , Johnson (1977) 481–87. 156/448 adjacent groups. 517 This network model of states and empires has implications for the means by which states expand and are governed. The network model invites us to consider the expansion of states by network influence rather than by hostile node takeover, that is, warfare. 518 Smith argues that state formation occurs by resource capture and selective boundary defence. 519 She uses the Inka, the Sasanian and the Mauryan empires as case studies to shows how ancient empires could govern large areas of space using a comparatively small number of network links and expand their control by integrating other local networks into their broader system of control. 520 Scholars have applied this network approach to other Near Eastern states. Karl Butzer has conceptualised the Early Bronze Age world between Spain and the Indus valley as a single politico- economic network comprised of a variety of local networks linked by flows of trade, production, redistribution, technology and institutions. 521 He describes its collapse in the late third millennium as a consequence of disruption of that global network through local and regional warfare and unrest. At a smaller scale, Leon Marfoe has argued that the small cities and villages of the Bronze Age Bekaa Valley (in modern Lebanon) were structured around loose networks of kinship groups fluctuating between sedentary dry farming practices and sheep and goat pastoralism rather than as consolidated territorial units. 522 Mario Liverani has argued that the Assyrian empire expanded through control of strategic points to become a network empire, then consolidated through expansion of administrative and military networks into a territorial empire. 523 At an early stage, he described the Assyrian empire as “not a spread of land but a 517 Shaw (1986). 518 Smith (2005) 845; Sinopoli (1994) 162–69. 519 Smith (2005). 520 Smith (2005) 839–44; Smith (2007) 31. 521 Butzer (1997); in particular, see Figure 1, p.284 for a schematic diagram of that Early Bronze W orld-Economy. 522 Marfoe (1979) esp. 16–17, 34–35. Marfoe discusses the possibility of similar situations existing in the region during the Hellenistic and early Roman periods (pp.19, 24-25). In the late-Hellenistic and early Roman period, we might imagine kingdoms like Emesa exhibiting similar structures. 523 Liverani (1988). 157/448 network of communications over which material goods are carried.” 524 Liverani describes the consolidation from network empire to territorial empire as the development of a “network whose mesh thickens”. 525 Liverani’s language shows that even when he speaks of a “territorial empire”, he is thinking of a network empire with a particularly “thick” or highly-linked network. In the Assyrian context, the consolidation of a network-empire was conducted with internal campaigns. The historical evidence suggests that similar dynamics were at play in Rome’s expansion into the Mesopotamian borderland. This chapter shows how those dynamics are visible in the Roman Imperial geographical writers specifically . Thinking in terms of networks is compatible with simultaneously thinking in terms of boundaries. Every network link can be conceptualised as a boundary crossing between space, gender, status, or any number of other potential boundaries that might separate two agents. Most such crossings will not be significant at any particular level of analysis, but some will be for any level that is chosen. Network Infrastructure of the Mesopotamian Borderland On the surface, the Roman Imperial geographical texts present a stable and coherent picture of Roman power and spatial control, but tension and dynamism underlie those narratives. For three centuries, the Romano-Parthian borderland in Mesopotamia was contested between two powerful and dynamic empires, mostly inconclusively . Until the second century , third parties in the borderland, be they minor kingdoms, cities or powerful individuals, influenced much of the interaction between Rome and Parthia. Large scale Roman military expeditions began at the beginning of the second century with Trajan’s invasion of Babylonia, but accelerated later in that century when Lucius Verus made lasting territorial conquests east of the Euphrates. Roman dysfunction and the rise of the Sasanid dynasty in the third century changed the inter-imperial dynamics of the space again, leading to a more rigidly defined and militarised border and a 524 Liverani (1988) 86. 525 Liverani (1988) 91. 158/448 succession of major wars resolved by official treaties and territorial exchanges. The geographical texts of the period show traces of these developments and the network interactions which underpinned them. In what follows, I will outline the historical context of two borderland kingdoms (Commagene and Osrhoene) and two cities (Nisibis and Palmyra) and their relation to the imperial powers in order to show how the descriptions of the Roman imperial geographic writers reflect changing political boundaries and borderland processes. At the appropriate point in the historical discussion, I will introduce the relevant geographic description. These studies illustrate the reflexive relationship between network interactions and geographical conditions in which each is capable of influencing the other. More significantly for my project, 159/448 Map 22: Sites in the Mesopotamian Borderland they show the varied ways in which those relationships were represented in geographical texts of the Roman Empire. Furthermore, they illustrate how geographical texts undercut and problematise relationships of political allegiance. The sites mentioned in these sections are located on Map 22: Sites in the Mesopotamian Borderland. Commagene in the Borderland Commagene existed as a political entity as early as the third century BCE. 526 According to Strabo, it was brought into the Seleucid Empire by Antiochus III. 527 The Commagenean kings claimed Persian links through the Armenian royal house and Macedonian links through Seleucid marriages and drew on both Iranian and Greek cultural traditions, including a combination of Greek and Iranian dynastic names (Antiochus, Samos, Mithridates). 528 Around 163/2 BCE as the Seleucid kingdom succumbed to internal and external pressures, the Seleucid governor of Commagene, Ptolemaios, rebelled from Antiochus IV Epiphanes, and Commagene became an independent Hellenistic kingdom. 529 It is likely that Tigranes of Armenia brought the kingdom under his rule when he conquered the remains of the Seleucid kingdom. 530 Tigranes 526 Cohen (2006) 30–32. 527 Sullivan (1977) 736–37. The Orontes named by Strabo (11.14.15) was king of Commagene. Memnon calls Antiochus III king of Syria and Commagene (᾽Αντίοχον τὸν Συρίας καὶ Κομμαγηνῆς καὶ ᾽Ιουδαίας βασιλέα), FgrHist 434 F 18.5; 18.9. 528 Sullivan (1977) esp. 736, 747. The artistic style of Commagene: Colledge (1987) 158–9; Cohen (2006) 30–32, esp. n.32. 529 Pressures on the Seleucids: Sullivan (1977) 737–47; Sherwin-White and Kuhrt (1993) 217–29. The rebellion of Ptolemaios: Diod. 31.19a; Sherwin-White (1993) 225–26; Sullivan (1990) 59–62; Sartre (2005) 23–24. The dating is from coin evidence for the start of a dynastic era: Sullivan (1977) 746. For the royal house and their connections to other Near Eastern royal houses, see the stemma in Sullivan (1977) 742. 530 The conquest is not attested in the literary sources, but inferred from the geography and from the presence of Armenian iconography in visual representations of Antiochus I: Sullivan (1977) 761. Appian (Syr., 48) reports that Tigranes ruled Syria for 14 years, Justin (40.1.4) gives 18 years. The discrepancy may reflect the time it took his Armenian armies to conquer the last of the Seleucid kingdom: Liebmann-Frankfort (1969) 193f.; Sullivan (1977) 760. Commagene was probably one of the first to be conquered because of its geographical position. Our only reference for conflict near Commagene is Strabo (16.2.3) who notes that Tigranes besieged the Seleucid queen in a city which was later attached to Commagene: “ἐν ᾧ τὴν Σελήνην ἐπικληθεῖσαν Κλεοπάτραν Τιγράνης ἀνεῖλε, καθείρξας χρόνον τινὰ ἡνίκα τῆς Συρίας ἐξέπεσεν.” (In [Seleucia (on the Euphrates)] Tigranes killed Selene, surnamed Cleopatra, who had been trapped there for some time when she was driven out of Syria.) This was Cleopatra Selene I, wife of Antiochus VIII Grypus, Antiochus IX Cyzicenus and Antiochus X Eusebes. Errington (2008) 276–77 provides a survey of the tumultuous final decades of the Seleucid dynasty. 160/448 probably simply made the Commagenean king his subject; there is no evidence for a break in the dynasty . Commagene’s position as a definitive borderland state in the early Roman period can seen in a brief survey of interaction between its kings and the Roman dynasts who jostled for power in Syria. Commagene appears sporadically in historical accounts of the first century CE. Its kings delicately managed a liminal borderland position and the shifting currents of late republican politics. Antiochus I of Commagene is named as one of the allies of Lucullus, but how that came about is unclear; Dio notes the friendship, but Pliny mentions a siege of Samosata by Lucullus. 531 Appian writes that Pompey forced (ἐπολέμησε) Antiochus to become his client (ἐς φιλίαν) then placed him in charge of “the parts of Mesopotamia which he [Pompey] had captured”. 532 Details on Pompey’s activities in northern Mesopotamia are slim, but this may have included the foothills as far east as Nisibis. Appian’s use of πολεμέω suggests that Antiochus resisted Pompey . If so, it may have been out of friendship to Lucullus, or out of a desire to retain a degree of independence which Lucullus may have granted him. However, Appian most likely overstated Antiochus’ resistance, either to glorify Pompey’s military achievements himself, or because he found Commagene listed among the locations over which Pompey celebrated his triumph and thus assumed that he had conquered the kingdom. 533 Pompey cannot have been overly concerned with the degree to which Antiochus resisted as he left him on the throne and extended his lands; for his part, Antiochus contributed troops to Pompey’s army at Pharsalus. 534 Most likely , Antiochus was one of the twelve foreign kings who offered their alliance to Pompey near the beginning of his campaign. Antiochus was an ally of Lucullus against Tigranes and maintained his 531 Dio 36.2.5; IGLSyr 52 = OGIS 404; Pliny NH 2.235. Plut. Luc. 29.5-6 does not list Commagene among Lucullus’ allies. 532 App. Mith. 106: “ὁ δὲ Πομπήιος καὶ τὸν Ταῦρον ὑπερελθὼν ἐπολέμησε μὲν Ἀντιόχῳ τῷ Κομμαγηνῷ, ἕως ἐς φιλίαν ὁ Ἀντίοχος αὐτῷ συνῆλθεν...” (After crossing the Taurus, Pompey fought Antiochus of Commagene until Antiochus entered into friendship with him). App. Mith. 114: “Ἀντιόχῳδὲ τῷ Κομμαγηνῷ Σελεύκειαν ἐπέτρεψε, καὶ ὅσα τῆς Μεσοποταμίας ἄλλα κατέδραμεν” (“[Pompey] entrusted Selucia and the other parts of Mesopotamia that he had overrun to Antiochus of Commagene.”). Strabo 16.2.3. 533 On Pompey’s triumph: Pliny NH 7.97-98; 37.12-19; Diod. 40.4. Pompey probably claimed to have conquered Commagene for the benefit of his reputation at Rome, Sullivan (1977) 763. 534 App. BC 2.49. 161/448 alliance with Rome. 535 Antiochus does not appear in the accounts of Crassus’s expedition. He seems to have remained loyal to Rome when the Parthians invaded Syria in 52 and 51 BCE; Cicero, then governor of Cilicia, reported to the senate that Antiochus sent him information about Parthian movements, although the probable movement of Parthian troops through Commagene aroused suspicion. 536 Antiochus was among the beneficiaries of Caesar’s clementia after Pharsalus. 537 However, the justification, or perhaps pretext as Dio notes, for the siege of Samosata by Ventidius and Antony in 39 BCE was that Antiochus was thought to have harboured Parthian fugitives after the defeat of Pakores’ invasion a few years earlier. 538 The Roman forces withdrew after Antiochus paid them a large sum. 539 Regardless of the truth of that accusation, Antiochus was active in both imperial centres: he appeared before the Senate to argue for his possession of a town near Zeugma; 540 and he married his daughter to a Parthian king and was subsequently killed in 36 BCE when he protested her murder by that king’s successor, possibly at court in Ctesiphon. 541 Antiochus’ son Mithridates II sent troops to help Antonius at Actium. 542 Octavian annexed Zeugma and Doliche from Commagene to Syria, perhaps because of this support. 543 Dio reports dynastic strife between Mithridates II and his brother 535 Plut. Pomp. 38.2; Sullivan (1977) 764. 536 Cic. Ad Fam. 15.1.2; 15.2.2; 15.3.1; 15.4.3. Sullivan (1977) 766–67; Syme (1995) 96. Dio (49.19.3) notes that the Parthians usually crossed the Euphrates at Zeugma, within Commagene. Caelius Rufus wrote to Cicero that Deiotarus had reported that the Parthians were moving through Commagene in 51 BCE (Cic. Ad Fam. 8.10.1). W e might well question whether Antiochus had the military capability to stop the Parthian forces had he wished to do so. Suspicion in 17 and 72 CE: Kennedy (1996a) 729–30. 537 Bell. Alex. 65: “Reges, tyrannos, dynastas provinciae finitimos, qui omnes ad eum concurrerant, receptos in fidem condicionibus impositis provinciae tuendae ac defendendae dimittit et sibi et populo Romano amicissimos.” (The kings, tyrants and rulers neighbouring the province, who had all come to him, he received into the fides of Rome on the condition that they watch out for and defend the province and sent them out as friends to himself and the Roman people). 538 Dio 49.20.3. Kennedy (1996a) 729. For accounts of Pakores’ invasion of Syria, see Justin 42.4.5-10, Josephus BJ 1.248-322. 539 Dio 49.20.5, 22.2; 48.41.5. Plut. Ant. 34 records the failure of V entidius and Antony to capture Samosata and Antiochus’ eventual payment of 300 talents. Josephus (BJ 1.321-22) reports that Herod assisted Antony in the successful capture of Samosata. Plutarch’s account seems more likely. 540 Cicero did not trust him: Ad Fam. 15.1.2. Cicero reports that he mocked and thwarted Antiochus, Ad Quint. Frat. 2.12.2. 541 Dio 49.23.3-4. The new king was Phraates IV , whose coup is also recorded by Plutarch (Ant. 37.1; Cras. 33.5). Plutarch does not mention the death of Antiochus. 542 Plut. Ant. 61. 543 The annexation of Zeugma in 31 BCE is deduced from the beginning of a new dating era on locally minted 162/448 Antiochus II in 29 BCE. 544 Although that brother is usually enumerated a king of Commagene, is it unclear whether he ever ruled; at most he held power jointly with Mithridates. Octavian clearly had no desire to see internal strife in Commagene; he summoned Antiochus II to Rome and had him executed. When Octavian visited the east in 20 BCE, he confirmed Mithridates III as king. 545 Around the same time, the young Mithridates was married to a princess of Atropatene. In these years leading up to the return of Crassus’ standards, relatively friendly relations between Rome and Parthia seem to have encouraged cross-border interaction between their respective client kingdoms. In the late republic, Commagenean kings forged new political links with a succession of Roman generals, local rulers and Parthian kings. They carefully balanced these network ties as best they could to ensure their own prosperity and continued rule. The political turbulence continued in the first century CE. The consolidation of the Roman Empire emphasised the relative importance of the network links between Rome and Commagene with the result that there were several administrative changes in Commagene as imperial policy fluctuated. Mithridates III was succeeded by Antiochus III, whose death in 17 CE set off a period of alternating direct and indirect Roman control of Commagene. According to Josephus, Antiochus’ death provoked conflict between the masses (τὸ πλῆθος) who wanted the monarchy to continue and the notables (τοὺς γνωρίμους) who wanted Commagene to be annexed by Rome. 546 Tacitus reports that the majority wanted Roman rule and the minority wanted the monarchy to continue (“plerisque Romanum, aliis regium imperium cupientibus”). 547 Josephus’ account is more probable; the nobility most likely saw more advantage in the removal of the monarchy , but Tacitus probably reports the official Roman version. 548 Tiberius dispatched Germanicus to coins, Wagner (1976) 64; Sullivan (1977) 780; Chaumont (1984) 69; Millar (1993) 29–30; Sartre (2005) 56. 544 Dio 52.43.1. Sullivan (1977) 728. 545 Dio 54.9.1-3; cf. Seut. Aug. 48. Sullivan (1977) 780. 546 Joseph. AJ 18.53-54. Kennedy (1996) 731. 547 Tac. Ann. 2.42.5. 548 Sullivan (1977) 784–85. After the final annexation of Commagene, the royal family appear in Rome in the senatorial class, Sullivan (1977) 795–97; Kennedy (1996a) 731. 163/448 settle the matter and Commagene was annexed to Syria in 18 CE. Tacitus reports that “Quintus Servaeus was appointed to Commagene, at that time first brought under the jurisdiction of a Praetor”. 549 Strabo wrote that the kingdom had recently (νῦν) become an eparchia. This word usually translated provincia, but perhaps here it indicates a prefecture subordinate to the governor of Syria as in the case of Judaea and the Decapolis. 550 Strabo’s use of νῦν cannot be used for precise dating, but here it must refer to Germanicus’ annexation, which took place only a few years before Strabo wrote book 16 of his Geography. 551 How was Commagene presented in Strabo’s text? 552 His description hints at Commagene’s turbulent journey through the waves of Seleucid contraction and Roman expansion: καθ’ ἕκαστα δὲ ἡ Κομμαγηνὴ μικρά τίς ἐστιν· ἔχει δ’ ἐρυμνὴν πόλιν Σαμόσατα ἐν ᾗ τὸ βασίλειον ὑπῆρχε, νῦν δ’ ἐπαρχία γέγονε· χώρα δὲ περίκειται σφόδρα εὐδαίμων, ὀλίγη [δέ]. Ἐνταῦθα δὲ νῦν ἐστι τὸ ζεῦγμα τοῦ Εὐφράτου· κατὰ τοῦτο δὲ Σελεύκεια ἵδρυται φρούριον τῆς Μεσοποταμίας προσωρισμένον ὑπὸ Πομπηίου τῷ Κομμαγηνῷ, ἐν ᾧ τὴν Σελήνην ἐπικληθεῖσαν Κλεοπάτραν Τιγράνης ἀνεῖλε, καθείρξας χρόνον τινὰ ἡνίκα τῆς Συρίας ἐξέπεσεν. 553 Clearly Strabo’s project would not allow him to present a picture of Commagenean political relations even as summary as that which I just gave. What does Strabo’s selective presentation tell us about his conception of 549 Tac. Ann. 2.56: “Commagenis Q. Servaeus praeponitur, tum primum ad ius praetoris translatis”. Kennedy (1996a) 712. 550 Strabo 16.2.3: “νῦν δ’ ἐπαρχία γέγονε”. Sartre (2005) 57. 551 Strabo’s note on the annexation of Commagene is one of the events by which Strabo’s Geography can be dated: Pothecary (1997); Dueck (1999); Pothecary (2002). 552 Strabo 16.2.2: Μέρη δ’ αὐτῆς τίθεμεν ἀπὸ τῆς Κιλικίας ἀρξάμενοι καὶ τοῦ Ἀμανοῦ τήν τε Κομμαγηνὴν καὶ τὴν Σελευκίδα καλουμένην τῆς Συρίας, ἔπειτα τὴν κοίλην Συρίαν, τελευταίαν δ’ ἐν μὲν τῇ παραλίᾳ τὴν Φοινίκην, ἐν δὲ τῇ μεσογαίᾳ τὴν Ἰουδαίαν. ἔνιοι δὲ τὴν Συρίαν ὅλην εἴς τε Κοιλοσύρους [καὶ Σύρους] καὶ Φοίνικας διελόντες τούτοις ἀναμεμῖχθαί φασι τέτταρα ἔθνη, Ἰουδαίους Ἰδουμαίους Γ αζαίους Ἀζωτίους, γεωργικοὺς μέν, ὡς τοὺς Σύρους καὶ Κοιλοσύρους, ἐμπορικοὺς δέ, ὡς τοὺς Φοίνικας. (W e set down as parts of Syria, beginning at Cilicia and Mt. Amanus, both Commagene and the Seleucis of Syria, as the latter is called; and then Coele-Syria, and last, on the seaboard, Phoenicia, and, in the interior, Judaea. Some writers divide Syria as a whole into Coelo-Syrians and Syrians and Phoenecians, and say that four other tribes are mixed up with these, namely , Judaeans, Idumaeans, Gazaeans, and Azotians, and that they are partly farmers as the Syrians and Coelo-Syrians, and partly merchants, as the Phoenicians.) Loeb trans. Radt (2002) 8.285–86. 553 Strabo 16.2.3: “In particular, Commagene is small. It has a fortified city , Samosata, in which lay the royal palace, but now it is a province. A small, but very fertile land lies around. In this place is now a bridge over the Euphrates (or Zeugma on the Euphrates). Opposite this lies Seleucia, a fortress of Mesopotamia, added to the dominion of Commagene by Pompey , in which Tigranes killed Selene, surnamed Cleopatra, who had been trapped there for some time when she was driven out of Syria.” Radt (2002) 8.286–88. 164/448 Commagene’s position as a borderland state? He notes the two most important sites of the area, the former royal capital, which is noted as such, and the crossing at Zeugma. Both were crossings of the Euphrates, critical for regional interactions between Syria and northern Mesopotamia, and astride routes of even greater long-range significance. Both of these crossings are fortified: Samosata is a fortified city (ἐρυμνὴν πόλιν) and Seleucia a fortress (φρούριον). Strabo’s Commagene is small, very fertile, highly fortified, and located on important routes. Strabo does not make the strategic importance of the kingdom explicit, but he gives the impression that it would be difficult to conquer. Strabo emphasises this difficulty with his description of Seleucia in which Cleopatra I Selene resisted Tigranes, the destroyer of the Seleucid kingdom, for some time (χρόνον τινὰ). It does not seem to have presented any difficulty to Roman power, here in the guise of Pompey . The addition of that fortress to the kingdom hints at the strategic importance of the kingdom. Pompey’s motive could have been to strengthen the kingdom he had just conquered or to place the fortress in the hands of an strong ally . Either motive suggests a concern for the defence of the region. So too does the annexation of the kingdom. Strabo presents a fortified area of implied strategic value, fertile and controlling important river crossings. As we shall see below , the wealth and strategic importance of Commagene are two of the motives which emerge for the Roman annexation of the kingdom. Strabo neither mentions nor alludes to the political divisions in Commagene between the parties who favoured the royal succession or Roman annexation. Nor does he mention the recurring concern for the loyalty of the Commagenean kings, who were accused of collaborating with Parthia whenever the kingdom was annexed or reduced in size, but especially after the final annexation of 72 CE. Importantly , Strabo presents the region as politically united and part of the Roman Empire. Caligula reestablished the kingdom under Antiochus IV in 37 CE. 554 The new king’s dominions were extended to include the Cilician coast. 555 After only a few years Caligula deposed him, but he was 554 After Germanicus’ annexation of Commagene, the children of Antiochus III were raised at Rome. 555 Restored by Caligula: Dio 59.8.2; Seut. Calig. 16.3. 165/448 restored again by Claudius. 556 Finally , after a long and active career of diplomacy and warfare, Antiochus was accused of plotting with the Parthians and deposed by Vespasian in 72 CE. 557 Antiochus IV attained Roman citizenship and retired peacefully to Greece, where his descendants were prominent. His sons attempted to resist the annexation, were defeated and fled to the Parthians who welcomed them, before finally returning to Roman territory and their father. V espasian’s final annexation of Commagene is relatively well reported in the sources. There has been much ancient and modern consideration of his motives for doing so, in part because it was a sudden reversal in fortune for Antiochus IV . Three related motives are discussed: the wealth of Commagene, the strategic importance of the area, and suspicion at links with Parthia. Our sources consistently paint Commagene as a wealthy region. Strabo describes the region around Samosata as particularly fertile (σφόδρα εὐδαίμων, ὀλίγη δέ). 558 In his description of Melitene, Strabo describes that region as similar to (παραπλησία) Commagene before describing fruit trees, olive oil and wine. 559 Regardless of its source, Commagene’s wealth was considerable. When V entidius and Antony attempted to punish Antiochus for supporting the Parthian invasions of the 40s BCE, the Commagenean king initially offered to pay off Antony with 1000 talents. 560 The offer was rebuffed; according to Plutarch, because Antony required a personal victory in the east, although Antony may have also thought that 556 Given back by Claudius, having been taken away by Caligula in 41 CE: Dio 60.8.1; Joseph. AJ 19.276. Wiedemann (1996) 223. 557 Career: Sullivan (1977) 787–91; Kennedy (1996a) 731. Accusations and annexation: Joseph. BJ 7.219-43; Suet. Vesp. 8.4. Sullivan (1977) 792–94; Kennedy (1996a) 729–31. 558 Strabo 16.2.3. Sullivan (1977) 734; Griffin (2000a) 27. 559 Strabo 12.2.1: “Ἔστι δ’ ἡ μὲν Μελιτηνὴ παραπλησία τῇ Κομμαγηνῇ· πᾶσα γάρ ἐστι τοῖς ἡμέροις δένδροις κατάφυτος, μόνη τῆς ἄλλης Καππαδοκίας, ὥστε καὶ ἔλαιον φέρειν καὶ τὸν Μοναρίτην οἶνον τοῖς Ἑλληνικοῖς ἐνάμιλλον.” (Melitene is similar to Commagene, for the whole of it is planted with fruit trees, the only country in all Cappadocia of which this is true, so that it produces, not only the olive, but also the Monarite wine, which rivals the Greek wines.) Loeb trans. 560 Plut. Ant. 34.3: “βουλόμενος ἕν γε τοῦτο τῶν ἔργων ἐπώνυμον αὑτοῦ γενέσθαι καὶ μὴ πάντα διὰ Οὐεντιδίου κατορθοῦσθαι.” ([Antony] insisted that this one exploit at least should bear his own name and that not all the successes should be due to V entidius.) Loeb trans. Kennedy calculates this as 24 million sesterces, Kennedy (1996b) 731. 166/448 Commagene was a richer prize that that. Tacitus described Antiochus IV as the richest of Rome’s client kings. 561 When Caligula restored the Commagenean dynasty in 37 CE, he repaid the revenues which Rome had gained over the previous 19 years; Seutonius reports that this amounted to 100 million sesterces. 562 The strategic importance of a rugged, fortified and defensible kingdom dominating north-south and east-west routes between Anatolia and the fertile crescent was apparent to ancient commentators as it is to modern scholars. 563 In his account of Vespasian’s annexation, Josephus records that Commagene’s location heightened concern at rumours of Parthian collaboration. 564 The strategic importance of Commagene had been demonstrated by Corbulo’s war against Parthia between 54 and 64 CE. As with most conflicts between Parthia and Rome, the prize and the site of conflict was Armenia, but the centre of Rome’s eastern power was in Syria. 565 Between 62 and 64 CE, the war threatened to spread south. 566 As governor of Syria, Corbulo massed on the Euphrates, threw up bridges and established a fortified bridgehead on the eastern bank of the Euphrates. 567 However, when he marched to the relief of Caesennius Paetus in Armenia, his route was on the Roman side of the river, through Commagene and Cappadocia, leaving a significant force to occupy those Euphrates fortifications (munimenta Euphrati imposita). 568 Commagene’s important geographical position both between the Roman province of Syria and the contested kingdom of Armenia, and astride the northern 561 Tac. Hist. 2.81.1: “Antiochus vetustis opibus ingens et servientium regum ditissimus” (Antiochus, who had vast ancient wealth and was the wealthiest of the subject kings). Josephus (BJ 5.461) also mentions Antiochus’ wealth. 562 Suet. Cal. 16.3. Thus Antiochus’ attempted payment to Antony amounted to about five years, above n.506. 563 Kennedy (1996a) 729–30 attributes the annexation of 17 CE to strategic concerns and that of 72 CE to collaboration with the Parthians. 564 Joseph. BJ 7.223-24: “ἔμελλε Καῖσαρ τοιούτου μηνύματος αὐτῷ προσπεσόντος μὴ περιορᾶν: καὶ γὰρ ἡ γειτνίασις τῶν βασιλέων ἐποίει τὸ πρᾶγμα μείζονος ἄξιον προνοίας: [224] τὰ γὰρ Σαμόσατα τῆς Κομμαγηνῆς μεγίστη πόλις κεῖται παρὰ τὸν Εὐφράτην, ὥστ᾽ εἶναι τοῖς Πάρθοις, εἴ τι τοιοῦτον διενενόηντο, ῥᾴστην μὲν τὴν διάβασιν βεβαίαν δὲ τὴν ὑποδοχήν.” (Now Caesar was disposed to take some care about the matter, since this discovery was made; for the neighborhood of the kingdoms made this affair worthy of greater regard; for Samosata, the capital of Commagene, lies upon Euphrates, and upon any such design could afford an easy passage over it to the Parthians, and could also afford them a secure reception.) 565 Corbulo’s campaign is described by Bivar (2000) 81–85. 566 V ologeses was based at Nisibis in 60 CE, Tac. Ann. 15.5. 567 Tac. Ann. 15.9; Dio 62.20. 568 Tac. Ann. 15.9-12; Dio 62.21. 167/448 route through northern Mesopotamia, was a compelling argument for bringing the kingdom under the direct control of the governor of Syria. 569 This would have been familiar to Vespasian from his extensive experience in the eastern provinces and kingdoms. 570 The proximity of the kingdom to Parthian territory cast suspicion on the loyalty of its rulers. Josephus reports that Vespasian’s annexation was precipitated by an accusation that Antiochus IV was planning to switch his allegiance to Parthia. 571 Antiochus had a long history of service to Rome, and the Flavian family in particular. He had contributed royal troops to Roman armies in Judea in 66, 67 and 70, and had supported V espasian in his imperial claim. 572 However, that his sons were received warmly by the Parthian king after their failed resistance to the annexation speaks to friendly ties to Parthia, or at least that the Parthian king was open to the idea of friendly relations and an alliance. Antiochus, like most of his predecessors, had maintained strong political connections with the important political actors of the Near East, so links with the Parthian empire would not be surprising. Despite violently annexing his kingdom, Vespasian restored relations with the deposed Commagenean house who entered senatorial politics in later generations. Commagene’s extensive network of regional political connections were valuable in the late 569 Importance of those routes for commerce and cultural contact: Sullivan (1977) 734, 762. 570 The annexation of Commagene was part of a broader reorganisation of the Roman military system in the east: Griffin (2000a) 38–40; Edwell (2008) 18–20 571 Joseph. BJ 7.219-22: “Ἤδη δὲ ἔτος τέταρτον Οὐεσπασιανοῦ διέποντος τὴν ἡγεμονίαν συνέβη τὸν βασιλέα τῆς Κομμαγηνῆς Ἀντίοχον μεγάλαις συμφοραῖς πανοικεσίᾳ περιπεσεῖν ἀπὸ τοιαύτης αἰτίας: [220] Καισέννιος Παῖτος, ὁ τῆς Συρίας ἡγεμὼν τότε καθεστηκώς, εἴτ᾽ οὖν ἀληθεύων εἴτε καὶ διὰ τὴν πρὸς Ἀντίοχον ἔχθραν, οὐ σφόδρα γὰρ τὸ σαφὲς ἠλέγχθη, [221] γράμματα πρὸς Καίσαρα διεπέμψατο, λέγων τὸν Ἀντίοχον μετὰ τοῦ παιδὸς Ἐπιφανοῦς διεγνωκέναι Ῥωμαίων ἀφίστασθαι συνθήκας πρὸς τὸν βασιλέα τῶν Πάρθων πεποιημένον: [222] δεῖν οὖν προκαταλαβεῖν αὐτούς, μὴ φθάσαντες τῶν πραγμάτων [ἄρξασθαι] πᾶσαν τὴν Ῥωμαίων ἀρχὴν πολέμῳ συνταράξωσιν.” (And now , in the fourth year of the reign of V espasian, it came to pass that Antiochus, the king of Commagene, with all his family , fell into very great calamities. The occasion was this: Cesennius Petus, who was governor of Syria at this time, whether it were done out of regard to truth, or whether out of hatred to Antiochus, (for which was the real motive was never thoroughly discovered,) sent an letter to Caesar, and therein told him that Antiochus, with his son Epiphanes, had resolved to rebel against the Romans, and had made a league with the king of Parthia to that purpose; that it was therefore fit to prevent them, lest they prevent us, and begin such a war as may cause a general disturbance in the Roman empire.) Loeb trans. 572 Troops: Joseph. BJ 2.500f (5000 in 66 CE); Joseph. BJ 3.68 (3000 in 67 CE); Joseph. BJ 5.460-63; Tac. Hist. 5.1.2 (in 70 CE). Support for Vespasian: Tac. Ann. 2.81.1. 168/448 republic, but may have ultimately harmed the dynasty in the new political environment dominated by two empires. In situations where two states or empires are in conflict, cross-border transactions between entities in their border zones are often discouraged or illegal. 573 Another factor in V espasian’s decision may have been the precedent that Rome had set. Roman emperors had spent much of the previous half-century manipulating peripheral client states to destabilise the Parthian monarchy . Following the destabilised political conditions of 69 CE, Vespasian might legitimately fear Parthian interference in Roman dynastic affairs and seek to shore up his power. Indeed, Vologeses had offered to support Vespasian in 69 and the Parthians supported false Neros in 79/80 and 88. 574 Commagene was strategically located, well-connected politically and wealthy . Regardless of the likelihood of rebellion from Roman control, such a rebellion would be potentially dangerous. Moreover, direct control of the richest client kingdom in the east would be lucrative. Pliny the Elder was a provincial official (probably procurator of Africa) when V espasian’s annexation of Commagene occurred. 575 He would have been aware of the general circumstances, if not Vespasian’s deliberations, when he wrote his geographical narrative of Syria. Nevertheless, his description of Commagene is cursory and he does not mention the annexation. 576 These factors are in keeping with Pliny’s general project: presenting the world as a static object of Roman power. In his initial overview of Syria, he lists Commagene among the places into which Syria used to be toponymically divided: “vocabatur... Commagene”. 577 Pliny’s use of the imperfect tense emphasises the past division of Syria by implicit contrast 573 See Minghi (1991) 16–17. 574 Tac. Hist. 2.82.3; 4.51.2. False Neros: Griffin (2000a) 41. 575 Syme (1969). 576 Pliny NH 5.85-86. 577 Pliny NH 5.66: “namque Palaestine vocabatur qua contingit Arabas, et Iudaea et Coele, dein Phoenice et qua recedit intus Damascena, ac magis etiamnum meridiana Babylonia, eadem Mesopotamia inter Euphraten et Tigrin quaque transit Taurum Sophene, citra vero eam Commagene et ultra Armeniam Adiabene, Assyria ante dicta, et ubi Ciliciam attingit Antiochia.” (For it was called Palestine where it touches the Arabians, and Judaea and Coele, then Phoenicia and further inland, Damascena, and then even further, southern Babylonia. It is called Mesopotamia between the Euphrates and Tigris rivers, Sophene across the Taurus. Then, Commagene on this 169/448 to the present unity of the area under Roman rule. Nevertheless, the term did not drop out of use; when his narrative journey around Syria reaches the Euphrates, Pliny uses “Commagene” to refer to the area as a sub- region of Syria. 578 However, Pliny shows no interest in Commagene as a region. In his Euphrates itinerary , Commagene is a term of topological specification without political connotations. It forms part of his unified and ordered Syria. With Vespasian’s annexation, Commagene ceased to become a borderland state, although it remained an identifiable sub-region within Syria. Ptolemy also used Commagene as a sub-region of Syria Coele. 579 By the beginning of the second century , the strategic importance of Commagene can be seen in the situation of a legionary base at the kingdom’s former capital (Σαμόσατα Λεγίων). In his description of the eastern limes, Ammianus calls the area Commagena in the first instance, glossing it with the more recent provincial appellation, nunc Euphratensis. 580 In fact, Ammianus’s inclusion of Hierapolis grants Commagena a broader application than the historical Commagene ever had; a consequence of the anachronistic implied equivalence of a contemporary and a historical space. The complex and turbulent history of Commagene’s experience in the Romano-Parthian borderland is reflected in the geographical descriptions of the early principate. Strabo’s more dynamic and historical geographic project refers to the different recent states of Commagenean independence in its description of the area. His description emphasises Roman control over the political landscape as a capstone on the history side of Sophene; Adiabene, previously called Assyria, on the other side of Armenia and Antiochia where it touches Cilicia.) 578 Pliny NH 5.85: “Arabiam inde laeva, Orroeon dictam regionem, trischoena mensura, dextraque Commagenen disterminat, pontis tamen, etiam ubi Taurum expugnat, patiens.” (By a distance of by three schoinoi, the Euphrates separates the region of Arabia called Osrhoene on the left and Commagene on the right, and allows a bridge, even where it smashes out of the Taurus.) 579 Ptol. Geog. 5.15.10-11. 580 Amm. Marc. 14.8.7: “Et prima post Osdroenam quam, ut dictum est, ab hac descriptione discrevimus, Commagena (nunc Euphratensis), clementer assurgit, Hierapoli (vetere Nino) et Samosata civitatibus amplis illustris.” (And first after Osdroene, which, as has been said, I have omitted from this account, Commagene, now called Euphratensis, gradually rises, famed for the great cities of Hierapolis (ancient Ninus) and Samosata.) 170/448 of the kingdom. For his part, Pliny nods to the historical divisions of the space in his listing of former names of the component parts of now-Roman Syria, but otherwise gives a static and ahistorical account which emphasises the present status of Commagene as a part of a stable imperial whole. While he shows the ambiguous status of northern Mesopotamia itself, Commagene is treated unambiguously as part of Roman Syria. Osrhoene in the Borderland On the eastern bank of the Euphrates, another independent state had appeared out of the dissolving Seleucid empire. The story of Osrhoene’s encounter with and eventual annexation by Rome is also reflected in the geographic sources, although less strongly than that of Commagene. Syriac chronicles of the sixth and eighth centuries CE date the foundation of the Abgarid dynasty in Edessa to the 130s BCE. 581 Because of the continued survival of Edessa’s Syriac name (Orhai), as well as the later explosion of Syriac literary culture itself, Ross suggests that the elite of Osrhoene were weakly Hellenised and prized its native culture. 582 The relationship between Osrhoene, the Arsacids and Tigranes is unclear, although there is no sign of a break in the Abgarid dynasty under either the Parthians or the Armenians. The Abgarids first appear in relation to Roman power when Pompey confirmed Abgar of Edessa on his way back to Syria from Armenia. 583 A few years earlier, Macedonian colonists at Carrhae had helped Afranius, one of Pompey’s generals, but there is no evidence that Carrhae was part of Abgar’s kingdom. 584 Roman power took hold in Syria through the co-option of pre-existing local networks and their existing power structures. 585 In 54 BCE, Crassus crossed the Euphrates, defeated the local Parthian satrap and 581 The Chronicle of Edessa and Chronicle of Zuqnin respectively: Ross (2001) 9; S. P . Brock (1979) 3–4 (Edessa); 10 (Zuqnin). 582 Ross (2001) 8. 583 Dio 40.20. Ross (2001) 10. The regnal numbering of the Abgarid kings is not secure and is largely omitted here. 584 Dio 37.5.5. Ross (2001) 10, 22–28. 585 Sartre (2005) 42–44. 171/448 occupied many cities between the Euphrates and the Balikh including Ichnae, Nicephorion, and Zenodotium. 586 Dio in particular stresses the willingness of Greek and Macedonian colonists to transfer their allegiance from the Parthians to the Romans. 587 There is probably an element of truth to this. Greek- speaking descendants of Greek and Macedonian colonists were likely to have widespread networks of communication with other such colonists in Syria and other areas under Roman control, and might hope for preferential treatment. However, the idea that local Greeks supported the invasion of a philhellenic Rome is a pro-Roman narrative. Our sources (Dio and Plutarch) were Greek writers within the Roman empire. The ideological appeal to a shared Greek culture was available to both Rome and local communities who were plausibly able to identify themselves as “Greek”. Moreover, it was not the exclusive property of Rome. The Parthian empire was also home to many communities founded by Macedonian dynasts, in particular at Seleucia on the Tigris. In the accounts of both Crassus and Dio, the Greek-named city Zenodotium with the Greek-named tyrant Apollonius was the only significant resistance to Crassus’ occupation of northwestern Mesopotamia; he fled to Parthia. Those Greek names do not necessarily indicate ethnicity , but they do denote an appeal to Greek culture which in this case clearly did not manifest itself as an allegiance to Rome. Nevertheless, such an appeal to network connections and influence on the basis of a shared ideology of Greekness appears frequently in the accounts of Rome’s northern Mesopotamian activities. The texts of the Roman geographic writers were part of the articulation and communication this narrative of a shared Hellenic culture. 586 Dio 40.12.2-13.4; Arr. FGrH 156 F 33; Plut. Cras. 17.2-4. On Crassus’ modest advance, barely past Osrhoene: Sherwin-White (1984) 282–3. Zenodotium: Olshausen, Eckart, “Zenodotium” BNP . Ichnae: Isidore 1: “῎Ιχναι, πόλις ῾Ελληνίς, Μακεδόνων κτίσμα.” (Ichnai, a Hellenic city and Macedonian foundation); Kessler, Karlheinz, “Ichnae” BNP . For Nikephorion see Appendix 1.12: Nikephorion. 587 Dio 40.13.1: “τῶν γὰρ Μακεδόνων καὶ τῶν ἄλλων τῶν συστρατευσάντων σφίσιν Ἑλλήνων ἄποικοι πολλοί, βίᾳ ἀχθόμενοι καὶ ἐς τοὺς Ῥωμαίους ὡς καὶ φιλέλληνας πολλὰ ἐλπίζοντες, οὐκ ἀκουσίως μεθίσταντο.” (For colonists in great numbers, descendants of the Macedonians and of the other Greeks who had campaigned in Asia with them, readily transferred their allegiance to the Romans, since they were oppressed by the violence of the barbarians, and placed strong hopes in the invaders, whom they regarded as friends of the Greeks.) Loeb trans. Local supporters of Crassus provided information about the approaching Parthian army: Plut. Cras. 18.2-3. 172/448 The development of links between Rome and Osrhoene is seen first in Crassus’ campaigns in 54 and 53 BCE. Abgar of Edessa was among the borderland rulers who supported Crassus. Abgar appears in Dio’s narrative as a master schemer who allies with Crassus in order to betray him on the battlefield. 588 In Plutarch’s account, an Arab phylarch named Ariamnes (probably to be identified with the same Abgar) 589 persuades Crassus to abandon the river bank and give the Parthians the advantage by advancing across the plains. 590 Both Dio and Plutarch are overly concerned to blame the loss on treachery rather than a tactical blunder perhaps born of arrogance. 591 However, Abgar’s actions are indicative of the kind of network links and power relations necessary for a small kingdom at a boundary between Roman and Parthian power. When a Roman army appeared on his doorstep and defeated the local Parthian forces, Abgar supported the 588 Dio 40.20.1-21.1: μέγιστον δὲ ὅμως αὐτοὺς ὁ Αὔγαρος ὁ Ὀρροηνὸς ἐλυμήνατο·ἔνσπονδος γὰρ τοῖς Ῥωμαίοις ἐπὶ τοῦ Πομπηίου γενόμενος ἀνθείλετο τὰ τοῦ βαρβάρου. καὶ τοῦτο μὲν καὶ ὁ Ἀλχαυδόνιος ὁ Ἀράβιος ἐποίησε. πρὸς γὰρ τὸ ἰσχυρὸν ἀεὶ μεθίστατο. [2] ἀλλ’ ἐκεῖνος μὲν ἐκ τοῦ προφανοῦς ἀπέστη, καὶ κατὰ τοῦτο οὐ δυσφύλακτος ἦν· ὁ δ’ Αὔγαρος ἐφρόνει μὲν τὰ τοῦ Πάρθου, ἐπλάττετο δὲ τῷ Κράσσῳ φιλικῶς ἔχειν, καὶ χρήματά τε ἀφειδῶς αὐτῷ ἀνήλισκε, καὶ τά τε βουλεύματα αὐτοῦ πάντα καὶ ἐμάνθανε καὶ ἐκείνῳ διήγγελλε, καὶ προσέτι εἰ μέν τι χρηστόν σφων ἦν, ἀπέτρεπεν αὐτόν, εἰ δ’ ἀσύμφορον, ἐπέσπερχε. [3] καὶ δὴ καὶ τοιόνδε τι τελευτῶν ἔπραξε. τοῦ γὰρ Κράσσου πρὸς Σελεύκειαν ὁρμῆσαι διανοουμένου, ὥστε ἐκεῖσέ τε ἀσφαλῶς παρά τε τὸν Εὐφράτην καὶ δι’ αὐτοῦ τῷ τε στρατῷ καὶ τοῖς ἐπιτηδείοις κομισθῆναι, καὶ μετ’ αὐτῶν (προσποιήσεσθαι γάρ σφας ἅτε καὶ Ἕλληνας ῥᾳδίως ἤλπιζεν) ἐπὶ Κτησιφῶντα μὴ χαλεπῶς περαιωθῆναι, [4] τούτου μὲν ὡς καὶ χρονίου ἐσομένου ἀμελῆσαι αὐτὸν ἐποίησε, τῷ δὲ δὴ Σουρήνᾳ ὡς καὶ ἐγγὺς [21.1] καὶ μετ’ ὀλίγων ὄντι συμμῖξαι ἔπεισε. καὶ μετὰ τοῦτο παρασκευάσας τὸν μὲν ὅπως ἀπόληται τὸν δ’ ὅπως κρατήσει (συνεχῶς γὰρ προφάσει κατασκοπῆς τῷ Σουρήνᾳ συνεγίγνετο), ἐξήγαγε τοὺς Ῥωμαίους ἀφροντιστοῦντας ὡς ἐπὶ νίκην ἕτοιμον, καὶ αὐτοῖς ἐν αὐτῷ τῷ ἔργῳ συνεπέθετο.” ([20.1] Nevertheless, the greatest injury was done them by Abgarus of Osroëne. For he had pledged himself to peace with the Romans in the time of Pompey , but now chose the side of the barbarians. The same was done by Alchaudonius, the Arabian, who always attached himself to the stronger party. [2] The latter, however, revolted openly , and hence was not hard to guard against; but Abgarus, while favouring the Parthian cause, pretended to be well disposed toward Crassus. He spent money for him unsparingly , learned all his plans and reported them to the foe, and further, if any of them was advantageous for the Romans, he tried to divert him from it, but if disadvantageous, urged him forward. [3] At last he was responsible for the following occurrence. Crassus was intending to advance to Seleucia so as to reach there safely with his army and provisions by proceeding along the banks of the Euphrates and on its stream; accompanied then by the people of that city , whom he hoped to win over easily , because they were Greeks, he would cross without difficulty to Ctesiphon. [4] Abgarus caused him to give up this course, on the ground that it would take a long time, and persuaded him to assail Surenas, because the latter was near by and had only a few men. [21.1] Then, when he had arranged matters so that the invader should perish and the other should conquer (for he was continually in the company of Surenas, on the pretext of spying), he led out the Romans in their heedlessness to what he represented as a victory in their very hands, and in the midst of the action joined in the attack against them.) Loeb trans. 589 Bivar (2000) 53; Ross (2001) 10. 590 Plut. Cras. 21.1-4. 591 Sherwin-White (1984) 288 blames the defeat on a tactical blunder. 173/448 victor until it was clear that the tide had turned at which time he returned his allegiance to the Parthians. 592 His choices must have seemed limited; the alternative path was shown by Apollonius of Zenodotium. The relatively limited appearance of Osrhoene and the Abgarids in the first century CE is reflected in Strabo’s narrative. In fact, while Strabo and Isidore both mention cities in the region, including Ichnae, Nikephorion, and Batnae/Anthemusia, Osrhoene itself is mentioned by neither. Given our poor knowledge of the extent of Abgar’s kingdom at in this period, only Strabo’s confused reference to Edessa can be said to illuminate his conception of Osrhoene, and even then, it is uncertain. That reference comes not in Strabo’s discussion of Mygdonia, which does include nearby Carrhae, but in his description of commercial traffic alongside the Euphrates. 593 Strabo’s confusion in applying the name “Edessa” to what is most likely a description of Hierapolis/Bambyke attests to the relative unimportance of Edessa and the Abgarid dynasty in his understanding of the borderland. The example of Abgar and Crassus would be repeated twice in the Julio-Claudian period. In 35 CE, a conspiracy of Parthian nobles successfully appealed to Tiberius to release a son of the former Parthian king, Phraates IV , held hostage at Rome. Tacitus reports that the Parthian king, Artabanus II, had become so brutal to his people that they would desert him for a king with a better claim to Arsacid descent and Roman support. 594 When the Parthian contender crossed the Euphrates, a number of important defectors joined him and he successfully captured Seleucia, but was driven out soon after by Artabanus’ new Iranian army . 595 In 49 592 After defeating the Romans at Carrhae, the Parthians recovered the cities Crassus had taken, Dio 40.28.1: “οἱ δὲ δὴ Πάρθοι τότε μὲν οὐ περαιτέρω τοῦ Εὐφράτου προεχώρησαν, ἀλλὰ τὴν ἐντὸς αὐτοῦ πᾶσαν ἀνεκτήσαντο.” (The Parthians at this time did not advance behind the Euphrates, but won back the whole country east of it.) Loeb trans. Bivar (2000) 55–6; Sherwin-White (1984) 290. 593 Strabo 16.1.27: “ὑπέρκειται δὲ τοῦ ποταμοῦ σχοίνους τέτταρας διέχουσα ἡ Βαμβύκη, ἣν καὶ Ἔδεσσαν καὶ Ἱερὰν πόλιν καλοῦσιν, ἐν ᾗ τιμῶσι τὴν Συρίαν θεὸν τὴν Ἀταργάτιν.” (Such travellers cross the Euphrates near Anthemusia, a place in Mesopotamia; and above the river, at a distance of four schoinoi, lies Bambycê, which is also called Edessa and Hierapolis, where the Syrian goddess Atargatis is worshipped.) For the geographical context of the passage, see Chapter 1; for a discussion of the Euphrates route, see Chapter 4. 594 Tac. Ann. 6.31. The first hostage released died in Syria (Tac. Ann. 6.32), but the plan continued with a second, Tiridates. 595 Initial success: Tac. Ann. 6.37. Ultimate failure: Tac. Ann. 6.42-4; Joseph. AJ 100. Tacitus (Ann. 6.32) reports 174/448 CE, another embassy of Parthian nobles attempted to place another pro-Roman candidate on the Parthian throne, again by crossing the Euphrates and marching on Ctesiphon with the support of local rulers. 596 According to Tacitus, this attempt was also a failure because of treachery on the part of another Abgar of Osrhoene, whom he calls Abgar rex arabum. In these accounts of Roman attempts to expand their influence into the Parthian empire in the first century CE, several kings east of the Euphrates, including Osrhoene, were seen as within the Parthian sphere, but generally pro-Roman and willing to support such ventures. These events reveal the presence and operation of connections between a network of conspirators nominally under Parthian rule (including the king of Osrhoene), 597 Parthian hostages at Rome, Roman administrators in Syria, and the Roman imperial court at Rome. These connections were not just personal and political, but involved the movement of Parthian nobles across the border and into the heart of the Roman empire. This is the background to Pliny’s description of the Mesopotamian borderland. Pliny’s description of Osrhoene does not mention these political and military manoeuvres but it does illuminate the kingdom’s critical and contested position in the inter-imperial borderland. Pliny’s description of Syria in book five and his description of northern Mesopotamia in book six overlap in northwest Mesopotamia. Pliny first mentions Osrhoene (Orroeon) in book five as an Arabian region on the opposite side of the Euphrates from Commagene. 598 Soon after he includes it in a list of areas in northern Mesopotamia (Osrhoene, praefectura Mesopotamiae, and an Arab tribe called Praetavi), within it he names Edessa and Carrhae. 599 In book six, Osrhoene appears again as an ethnic group (Orroei); one of a list of Arab intrigue by Artabanus against the pretender and his allies. 596 Tac. Ann. 12.10-14; Bivar (2000) 76–8; Ross (2001) 10; Edwell (2008) 16–17. 597 Abgar of Osrhoene was definitely involved in 49 CE. The route of the invasion directly through Osrhoene argues strongly for his involvement in 35 CE as well. 598 Pliny NH 5.85: “Arabiam inde laeva, Orroeon dictam regionem, trischoena mensura, dextraque Commagenen disterminat, pontis tamen, etiam ubi Taurum expugnat, patiens.” (By a distance of by three schoinoi, the Euphrates separates the region of Arabia called Osrhoene on the left and Commagene on the right, and allows a bridge, even where it smashes out of the Taurus.) 599 Pliny NH 5.86: “Arabia supra dicta habet oppida Edessam, quae quondam Antiochia dicebatur, Callirhoem, a fonte nominatam, Carrhas, Crassi clade nobile. iungitur praefectura Mesopotamiae, ab Assyriis originem trahens, in qua Anthemusia et Nicephorium oppida. mox Arabes qui Praetavi vocantur; horum caput Singara.” (Arabia, 175/448 tribes (Arabum gente) in Mesopotamia. 600 As I argue in Chapter 3, Pliny’s division of narrative space between books five and six corresponds to a division of imperial space between Roman and Parthian spheres. This overlap places Osrhoene/the Orroei, Anthemus(ia) and Nikephorion (all in the western part of northern Mesopotamia) in both Roman and Parthian space. 601 By doing so, Pliny figures the western part of the borderland between the Euphrates and the Balikh rivers as a contested space which exists within the overlapping spheres of Roman and Parthian influence. The contrast is emphasised by the comparative order and disorder of the two worlds which laid claim to the space. In 114, Trajan invaded and annexed Armenia. 602 Either in the same year or the next, he moved down from Armenia to the Parthian headquarters at Nisibis and conquered northern Mesopotamia. 603 The local rulers, including Abgar of Osrhoene, joined him or were compelled to do so. 604 In 116, having wintered in Antioch, Trajan invaded along the Euphrates route, seized Babylonia, received the submission of Attambelos V of Mesene (Characene) at the head of the Arabian gulf, and crowned a Parthian client king. 605 Trajan established two provinces: Armenia and Mesopotamia, but the conquered cities immediately revolted. Abgar above mentioned, has the towns of Edessa, formerly called Antiochia; Callirhoë, named from its fountain, and Carrhæ, famous for the defeat of Crassus. Adjoining to this is the praefecture of Mesopotamia, which derives its origin from the Assyrians, and in which are the towns of Anthemusia and Nicephorium. Next are the Arabians, called Praetavi whose capital is Singara.) 600 Pliny NH 6.117-18: “item in Arabum gente qui Orroei vocantur et Mandani Antiochiam quae a praefecto Mesopotamiae Nicanore condita Arabis vocatur. [118] Iunguntur his Arabes introrsus Eldamari, supra quos ad Pallacontam flumen Bura oppidum, Salmani et Masei Arabes.” (And in Arabia, the people called the Orroei, and Antiochia of the Mardani, founded by Nicanor, the governor of Mesopotamia, and called Arabis. [118] Joined to these in the interior are the Eldamari Arabs (beyond whom is the town of Bura near the river Pallaconta) and the Salmani and Masei Arabs.) 601 Orroei: NH 5.85, 6.117. Anthemus(ia) and Nikephorion: NH 5.86; 6.118. 602 Dio 68.19-20, Moses Khorenats'i 2.55; Bivar (2000) 87–88. 603 Lightfoot (1990) 117–20; Bivar (2000) 88; Ross (2001) 30–33. On the source and dating difficulties posed by Trajan’s campaign, see Griffin (2000b) 123–4. 604 Dio 68.18, 68.21. Others local rulers who joined Trajan included Mannus ὁ τῆς Ἀραβίας τῆς πλησιοχώρου (Dio 68.21-22. Probably Ma‘nu of Singhara, Ross (2001) 34); Sporakes, phylarch of Anthemusia (Dio 68.21); Manisaros (Of Gordyene? Dio 68.22.1: καὶ τοῦ Μανισάρου πρέσβεις ὑπὲρ. εἰρήνης διὰ τὸ τὸν Ὀρρόην ἐπιστρατεύειν αὐτῷ πέμψαντος καὶ τῆς τε Ἀρμενίας καὶ τῆς Μεσοποταμίας ἑαλωκυίας ἀποστῆναι ἑτοίμως ἔχοντος); Mebarsapes, king of Adiabene (Dio 68.22.2: καὶ τὸν Μάννον ὑπώπτευεν ἄλλως τε καὶ ὅτι συμμαχίαν Μηβαρσάπῃ τῷ τῆς Ἀδιαβηνῆς βασιλεῖ πέμψας πᾶσαν αὐτὴν ὑπὸ τῶν Ῥωμαίων ἀπεβεβλήκει). 605 Dio 68.26.4-30. Lightfoot (1990) 120–1; Bivar (2000) 89–90. 176/448 may have been reluctant to support the Romans. He had delayed meeting Trajan until Roman forces controlled all of his neighbours and Trajan himself was marching towards Edessa. 606 Edessa was among the cities that rebelled in 116; it was captured and burned when the rebellion was crushed. 607 Abgar died in those rebellions and for two years, Edessa was occupied by Roman troops. 608 The Parthian king Osroes was able to recover his lost territory , although the Parthian king crowned by Trajan may have maintained control of the Mesopotamian borderland for several more years. 609 Trajan himself apparently planned a further campaign to recover the territory , but he fell ill and died while returning to Rome. There are no extant geographical descriptions of Trajan’s short-lived province of Mesopotamia. While Ptolemy represents the world at around this time, his Mesopotamia is a geographical region, not an administrative area. Despite sharing a name, the two spaces were not collocated. Ptolemy’s geographical region was bounded by the Taurus, the Tigris and Euphrates and the Naarmalcha canal. Trajan’s province probably included the northern Mesopotamian shelf from the border of Osrhoene to the Tigris and down the Khabur river to the Euphrates. It would have definitely included Nisibis and Singara, and perhaps Amida and Constantina. A Roman milestone dating to 116 CE was found near Singara on the road to Nisibis. 610 Epigraphic and numismatic evidence shows that Dura Europos also was briefly under Roman control, but it 606 Dio 68.21; Ross (2001) 34; Millar (1993) 102. 607 Dio 68.30.1-2: “μαθὼν δὲ ταῦτα τόν τε Λούσιον καὶ τὸν Μάξιμον [2] ἐπὶ τοὺς ἀφεστηκότας ἔπεμψε. καὶ οὗτος μὲν ἀπέθανεν ἡττηθεὶς μάχῃ, Λούσιος δὲ ἄλλα τε πολλὰ κατώρθωσε καὶ τὴν Νίσιβιν ἀνέλαβε, τήν τε Ἔδεσσαν 3 ἐξεπολιόρκησε καὶ διέφθειρε καὶ ἐνέπρησεν. ἑάλω δὲ καὶ ἡ Σελεύκεια πρός τε Ἐρυκίου Κλάρου καὶ πρὸς Ἰουλίου Ἀλεξάνδρου [3] ὑποστρατήγων, καὶ ἐκαύθη.” (When [Trajan] learned of the revolt, he sent Lusius and Maximus against the rebels. [2] The latter was defeated in battle and perished; but Lucius, in addition to many other successes, recovered Nisibis, and besieged and captured Edessa, which he sacked and burned. Seleucia was also captured by Erucius Clarus and Julius Alexander, lieutenants, and was burned.) Loeb trans. 608 Ross thinks it more likely that he died defending his new alliance with Trajan than opposing the still powerful Roman forces in the area. Trajan’s Parthian appointee retaining control in northern Mesopotamia: Ross (2001) 34–35; Bivar (2000) 92–3. 609 Trajan’s provinces: Lightfoot (1990) 121–4. Rebellion: Lightfoot (1990) 124; Bivar (2000) 90–1. Despite an attempt on his return, Trajan was never able to capture Hatra, Bivar (2000) 91–2; Birley (2000) 134. Judgement of Trajan’s success: Ball (2000) 16. 610 AE (1927), no. 161. 177/448 is not clear whether it was under the jurisdiction of the governor of Syria or Mesopotamia. 611 When it was conquered later in the second century it was attached to Syria. 612 The southeast boundary of the province may have been the territory of Hatra, which Trajan failed to conquer. Osrhoene also remained an independent client kingdom beyond the direct administration of a Roman governor. For the next fifty years, factions connected to Roman and Parthian networks squabbled for control of Osrhoene. 613 A succession of local kings backed by pro-Roman or pro-Iranian parties ruled the kingdom until the capture of Edessa by Lucius Verus. Numismatic evidence shows a king Ma‘nu who identifies himself as ΜΑΝΝΟΣ ΦΙΛΟΡΟΜΑ[ΙΟΣ] minting coins with Roman figures and Greek legends as well as a king Wael minting coins with Parthian figures and Syriac legends. 614 These kings must have had networks of contact and support within the respective empires. The Chronicle of Zuqnin relates that Ma‘nu was driven out of Edessa and into the Roman empire, probably by a Parthian intervention that placed Wael on the throne. 615 Ma‘nu seems to have regained his throne when Lucius Verus invaded. 616 Lucius Verus managed his war against Parthia from Antioch. 617 In 165 or 166, he sent Avidius Cassius through Osrhoene where he restored Ma‘nu Φιλορομαιος to Edessa, down the Euphrates where he recaptured Dura-Europos and other Euphrates fortresses and into Babylonia where he sacked Ctesiphon. 618 611 Edwell (2008) 115; Millar (1993) 102; Hopkins (1979) 68. A document found at Dura-Europos shows that the city was back in Parthian hands by 121 CE. 612 Millar (1993) 102. Probably because of strong pre-exisiting connections to Palmyra. On Trajan’s campaign opening the Euphrates to Palmyrene influence, see Edwell (2008) 21. 613 Ross (2001) 36–7. 614 Coins depicting V ologaeses IV on the obverse and “Wael the King” (ܥܟܠܡ ܠܥܘ) on the reverse: Hill (1965) 91– 92, nos. 1–3 (Wael); Ross (2001) 36; Millar (1993) 112. Coins either reading “Ma‘nu the King” (ܥܟܠܡ ܘܢܥܡ) or depicting members of the Antonine dynasty and marked ΜΑΝΝΟΣ ΦΙΛΟΡΟΜΑ[ΙΟΣ]: Hill (1965) 92–93, nos. 5–9 (Ma‘nu); Ross (2001) 36; Millar (1993) 112. 615 Ross (2001) 36. 616 BMC Arabia p.xcvii. 617 Dio 71.1.3; SHA Marcus 8.9; V erus 5.8. On the novelty of the move: Millar (1993) 111–12. 618 Dio 71.2.2; Edwell (2008) 23–26; Bivar (2000) 93. Edessa: Dio 71.2.1. Ctesiphon: Dio 71.2.3; but note Ball's caution: “The ease with which the Romans claimed to have taken Ctesiphon virtually as a matter of course during so many of their Parthian campaigns, but without and long-term effect, makes one suspect the sources somewhat, or even whether it was actually taken”, (2000) 17, n.47. 178/448 Roman control now seems to have extended to a line roughly from the Jebel Sinjar down the Khabur to Dura Europus. 619 This line was a route following a fertile strip between a major administrative and population centre in the north (Nisibis) and an important fortress and Palmyrene trading post in the south (Dura-Europos). The political organisation of the new territory is unclear. 620 W e know that Dura was added to the province of Syria. 621 Osrhoene was still held by the Abgarids of Edessa, although Roman influence had increased and probably included permanent garrisons. 622 A collection of inscriptions found at Sumatar Harabesi near Edessa seem to show a new religious centre aimed at extending Edessan control over the ‘Arab region to the east of the kingdom. 623 Verus installed an unknown number of garrisons from Osrhoene to Adiabene, including one at Nisibis, but there is no evidence that Lucius Verus restored the Trajanic province of Mesopotamia. Evidence for these garrisons is indirect. The kings of Edessa and Adiabene defended their attack on Nisibis in 194 on the grounds that the Roman soldiers there had supported Pescennius Niger, then governor of Syria. 624 Moreover, Dio reports that the kings of Osrhoene and Adiabene demanded the removal of other Roman garrisons which existed in their territories. 625 Ross calls the Roman presence in northern Mesopotamia a state of de facto annexation. 626 However, the scarcity of administrative details are a problem. If there was a basis in truth to the claim that the garrison in Nisibis supported Niger, then we might expect that these garrisons fell under his authority as governor of Syria. However, they may have been organised separately under an officer who sided with Niger. Nor do we know their precise purpose, activities, or what influence they had, or were intended to have on the cities in which they were stationed. Our only 619 Bivar (2000) 94. 620 Ross (2001) 37–9; Millar (1993) 113. 621 See Appendix 1.8: Dura-Europus. 622 Dio 75.1.3: “οὐ μέντοι οὔτε τὰ τείχη ἃ ᾑρήκεσαν ἐκλιπεῖν οὔτε φρουροὺς λαβεῖν ἤθελον, ἀλλὰ καὶ τοὺς λοιποὺς ἐξαχθῆναι ἐκ τῆς χώρας ἠξίουν.” (Yet they were unwilling either to abandon the forts that they had captured or to receive garrisons, but actually demanded the removal from their country of such garrisons as still remained.) Loeb trans. 623 Ross (2001) 39–43; Drijvers (1972). 624 Dio 75.1.2-3. 625 Dio 75.1.3, see above p.153. 626 Ross (2001) 38. 179/448 geographical source from this time is the Antonine Itinerary. This document (or collection of documents) ends in the east with a dense cluster of routes connecting Edessa to Syria. 627 Neither Nisibis nor any sites in Osrhoene further east than Edessa appear in the text. The Itinerary certainly gives the impression that Edessa and Osrhoene were a well-integrated part of the Roman Near East. The reign of Septimius Severus saw the end of Osrhoene as a borderland kingdom and the establishment of the provincial organisation which would remain until the end of our period. Severus disregarded the claim by Osrhoene and Adiabene that they had been acting in his interests in attacking Nisibis in 194 and invaded northern Mesopotamia immediately after his victory over Niger at Issus in the same year. 628 The Parthians seem not to have been involved in this first of Septimus Severus’ eastern campaigns. 629 Severus crossed the Euphrates in 197/198, marching at least as far south as Ctesiphon, which he sacked, and as far east as Hatra, which he did not. 630 As Trajan had done, Severus now organised northern Mesopotamia into the Roman provincial framework in the form of two new provinces called Osrhoena and Mesopotamia. The former comprised much of the territory of Abgarid Osrhoene, although Abgar VIII did not lose his throne. Edessa and a small territory became a client kingdom. A milestone found 48 milia passum to the west and the presence of the colonia established at Carrhae to the south give a general sense of the size of the new kingdom. 631 The rest of Osrhoene became a province under an equestrian procurator. 632 Further east, the province of Mesopotamia, under an equestrian praefectus, was probably similar in area to the Trajanic province and included the new coloniae of Nisibis, Singara and Reshaina. 633 In 212/213, 627 The Antonine Itinerary and this Edessan network is discussed in Chapter 5. 628 Ross (2001) 47–8. 629 Dio reports that they were dissuaded by bad omens (75.3.1) but internal issues of some sort seem more likely , Bivar (2000) 94. 630 Millar (1993) 121; Edwell (2008) 26; Ball (2000) 18. Ball suggests that the ease of seizing Ctesiphon was because Septimus had launched a surprise attack during a truce. 631 Ross (2001) 50–1. 632 Millar (1993) 125. 633 Millar (1993) 125–6, 143–44. 180/448 Caracalla abolished the Abgarid kingdom, incorporated it into the province of Osrhoena as a colonia. 634 The provinces of Osrhoena and Mesopotamia suffered several Sasanian invasions over the subsequent centuries, but the division of the geographical area of northern Mesopotamia between the two administrative regions lasted until the time of the Exposito and Ammianus. Nisibis Nisibis appears infrequently but regularly in the historical record until the 3 rd century when the expansion of the Roman frontier encompassed the city . The importance of the city within Mesopotamia is indicated by the frequency with which it appears in the geographical texts. This regional importance was no doubt the reason for its refoundation as Mygdonian Antioch by the Macedonians, a refounding which both Strabo and Pliny note. Ptolemy’s omission of this alternate name may indicate that it had dropped out of use by the second century . However, Ptolemy prefers to use local names for sites and seldom records multiple names for a single site unless there is doubt about the name, so his silence on the Macedonian name of Nisibis is not informative. While we know that Nisibis was a major city of the second-century provinces of Mesopotamia, we have no good information on the bounds of those provinces. I use the plural “provinces” because we do not even know if Roman administrators conceived of subsequent iterations of Mesopotamia as re-establishments of a previous province or as entirely new provinces with no link to the administrative history of the area. Our first geographical attestation of Nisibis after the establishment of those provinces is that of the Expositio Totius Mundi. Post hos nostra terra est. Sequitur enim Mesopotamia et Osdroena. Mesopotamia quidem habet 634 Dio 77.12.1ª-1². P .Dura 1 of 243 CE notes the “liberation of Antoniana Edessa Colonia Metropolis: 31 years prior, that is, in 213 (Drijvers and Healey (1999) 240; Sartre (2005) 344, with n.6). This refers to the annexation and probably the status of colonia as well; metopolis may have come later. By the mid-third century , Edessa had become “Edessa Antoniana Colonia Metropolis Aurelia Alexandria”; Ross ((2001) 57–63, esp. 59) argues that these titles were added in stages. Millar (1993) 144. 181/448 civitates multas et varias, quarum excellentes sunt quas volo dicere. Sunt ergo Nisibis et <Amida>, quae in omnibus viros habent optimos et in negotio valde acutos et bene <vendentes>. Praecipue et divites et omnibus bonis ornati sunt: accipientes enim a Persis ipsi in omnem terram Romanorum vendentes et ementes iterum tradunt, extra aeramen et ferrum, quia non licet hostibus dare aeramen aut ferrum. Istae autem civitates semper stantes deorum et imperatoris sapientia, habentes moenia inclita, bello semper virtutem Persarum dissolvunt; ferventes negotiis et tranigentes cum omni provincia bene. Deinde Osdroenae Edessa et ipsa civitas splendida. 635 The celebration of the impregnability of the walls of Nisibis and Amida places the composition of the text before the destruction of Amida in 359/360 and the surrender of Nisibis to the Persians in 363. 636 The Expositio represents the area of the Mesopotamian borderland as divided between these two provinces, Osrhoena and Mesopotamia. Like many of our sources for the fourth century , it does not provide much detail on the boundaries of those administrative spaces. However, this text does mention Nisibis’ role as a place of exchange between the two empires, perhaps its most famous role in a Roman context: “for they receive sellers from Persia into the whole Roman empire and send back buyers” (“accipientes enim a Persis ipsi in omnem terram Romanorum vendentes et ementes iterum tradunt”). 637 This was a legal status mandated by the treaty of 298 between Diocletian and Narsē. That treaty elevated Nisibis to a focal role in inter-imperial military , diplomatic and economic activity and the borderland processes it produced were reflected in the geographical descriptions of the city . Perhaps the most important clause of the treaty of 298 is the notice of the status of Nisibis as the point of commercial contact between Roman and Persian space. The exact ramifications of this attempt to control movement and commercial activity in the borderland by legal means are unclear. A summary of the 635 Expositio 22: “After this is our land; for Mesopotamia and Osrhene follow . Indeed, Mesopotamia has many diverse cities; I wish to speak of those which are outstanding. Namely , Nisibis and Edessa, which have men who are the best men in all things, very acute in business, and good salesmen. The cities are especially wealthy and supplied with all goods: for they receive sellers from Persia into the whole Roman empire and send back buyers. Except bronze and iron, since it is not permitted to give bronze or iron to foreigners. But those cities ever- standing by means of the forethought of the gods and the emperor, having famous walls, always destroy the courage of the Persians in war. Fervent in business and dealing well with every province. Then Edessa of Osdroenae also a very splendid city.” See Chapter 5 for a discussion of the amendments Amida and vendentes. 636 See Introduction. On the image of an impregnable Nisibis, see also Amm. Marc. 25.9.8. In the first century BCE, Nisibis’ strong walls caused Lucullus difficulties, Dio 36.6-7. 637 Peter the Patrician (fr.13-14). Dignas and Winter (2007) 122–130. 182/448 treaty is transmitted to us by the sixth century Byzantine historian Peter the Patrician. Οτι Γαλεριος και Διοκλητιανὸς εἰς Νισιβιν συνηλθον, ἔνθα κοινῇ βουλευσάμενοι στελλουσιν εἰς Περσιδα πρεσβευτην Σικοριον ΠροΒον ἀντιγραφεα της μνημης... ην δὲ τα κεφάλαια της πρεσβειας ταῦτα, ωστε κατα τὸ ἀνατολικὸν κλιμα την Ἰντηληνην μετα Σοφηνης και Ἀρζανηνην μετα Καρδουηνῶν και Ζαβδικηνης Ῥωμαιους ἔχειν, και τὸν Τιγριν ποταμὸν ἑκατερας πολιτειας ὁροθεσιον ειναι, Ἀρμενιαν δὲ Ζινθα τὸ κάστρον εν μεθοριῳ της Μηδικης κειμενον ὁριζειν, τὸν δὲ Ιβηριας βασιλεα της οἰκειας βασιλειας τα συμβολα Ῥωμαιοις ὀφειλειν, ειναι δὲ τοπον τῶν συναλλαγμάτων Νισιβιν την πολιν παρακειμενην τῷ Τιγριδι. τουτων ὁ Ναρσαῖος ἀκουσας, επειδη πρὸς μηδὲν τουτων ἡ παροῦσα τυχη ἀρνεῖσθαι αὐτὸν συνεχωρει, συνεθετο τουτοις απασι· πλην ινα μη δοξῃ ἀνάγχῃ πάντα ποιεῖν, διηρνησατο μονον τὸ τοπον εἶναι τῶν συναλλαγμάτων την Νισιβιν. ὁ δὲ Σικοριος· "ενδεδωκεναι χρη πρὸς τοῦτο. ουτε γαρ αὐτοκράτωρ ἡ πρεσβεια, και περι τουτου εκ τῶν αὐτοκρατορων οὐδὲν επετετραπτο." τουτων ουν συντεθεντων... 638 The main points of the treaty relevant to this discussion are: the transfer of the Transtigritanian provinces to Roman control; 639 the declaration of the Tigris as a boundary (ὁροθεσιον) between Roman and Persian space; and the requirement that Nisibis be the place where contacts are arranged and witnessed between the two empires (τοπον τῶν συναλλαγμάτων). It is notable that the only clause which Narsē attempted to reject (unsuccessfully) was that concerning Nisibis. The treaty of 298 changed the legal status of Nisibis, but the exact nature of the change is not entirely clear. Petrus reports that Nisibis was to become the “place for transactions” (τοπον τῶν συναλλαγμάτων). This is usually taken to mean that Nisibis became the only site at which trade between Rome and Persia could occur. Dignas and Winter, for example, write of Nisibis becoming “a huge trans- 638 Petrus Patricus, Fr 14: “When Galerius and Diocletian met in Nisibis, they deliberated together and sent into Persia the archivist [ἀντιγραφεα της μνημης] Sicorius Probus as an ambassador... These were the main points: that in the eastern region, the Romans should have Intelene with Sophene and Arzanene with Karduene and Zabdikene, that the Tigris river should be the boundary [ὁροθεσιον] between the two states, that the fortress Zintha which lies on the edge of Media should be the boundary with Armenia, that the king of Iberia should owe the tokens of his royal family to the Romans, and that the city of Nisibis which lies beside the Tigris should be the place of contracts [τοπον τῶν συναλλαγμάτων]. When Narsē heard these terms (since his present fortunes did not allow him to refuse any of them) he agreed with all of them, except, so that he didn’t seem forced to agree entirely , denied only that Nisibis should be the place of contracts. And Sicorius said “You must agree with this, for the embassy is not the emperor and nothing concerning this point has been entrusted to us by the emperor.” So Narsē agreed with them...” 639 Ammianus (25.7.9) calls them the regiones Transtigritanas. 183/448 shipment centre”. 640 However, A. D . Lee notes several difficulties in taking this law as requiring all trade to pass through Nisibis. 641 In particular, several fourth century sources note the trade wealth of other border cities. Ammianus mentions that Callinicum at the junction of the Euphrates and the Balikh was a city of rich trade (commercandi opimitate). 642 The Expositio uses plural verbal forms to describe the trading wealth and trade restrictions of Nisibis and Amida (ornati sunt, accipientes, ementes, tradunt). It is also possible that the plurality should apply beyond those two cities: Nisibis and Amida are outstanding (excellentes) among the cities of Mesopotamia. 643 Lee points out that if the treaty of 298 restricted all movement to Nisibis and the treaty of 363 (see below) retained the terms of the first treaty , then it is odd that an imperial edict of 408/9 allows trade at three locations (Nisibis, Callinicum, and Artaxata) while giving the aim as preventing spying. 644 In the singular, a συνάλλαγμα is a contract. This clause may have made Nisibis the place at which contracts for inter-imperial trade had to be registered rather than the place through which goods had to move. Whether Nisibis became a transshipment centre or a contract depository , two questions remain. Why Nisibis? And why Nisibis alone? Restricting trade between the Roman and Persian empires to only one site would have had several major effects and would have caused major difficulties, both for merchants and officials. Most importantly , it located the primary commercial site between the two empires in Roman territory . Channelling all cross- border trade through a single city would have made it easier for Roman administrators to collect taxes on that trade. 645 The Sasanians could still impose their own taxes elsewhere, but maintaining multiple locations placed a greater logistical burden on them. Moreover, it allowed trans-border movement more generally to be 640 Dignas (2007) 197. 641 Lee (1993) 62–64. 642 Amm. Marc. 23.3.7. 643 Expositio 22: “Mesopotamia quidem habet civitates multas et varias, quarum excellentes sunt quas volo dicere.” Translated above, p.153. 644 CJ 4.63.4 (“ne alieni regni, quod non convenit, scrutentur arcana”), see below p.152. Lee (1993) 63. 645 For Roman customs practices in Asia, see Cottier et al. (2008). For customs duties and tax in Egyptian trade: Young (2001) 66–69. 184/448 controlled. The explicit notification that the intention of the edict of 408/9 was to prevent the flow of sensitive information across the border suggests that the same considerations were behind the restrictions of 298 and 363 as well. These administrative and strategic benefits to the Roman Empire were balanced by costs to private costs to merchants. The two empires shared a border from the Black Sea to Arabia, as well as the potential for maritime connections around the Arabian peninsula. The further an inter-imperial merchant operated from Nisibis, the more difficult and expensive observing this treaty requirement would have been. Consider the burden on merchants operating along the Black Sea coast. Such a restriction would both be difficult to enforce, and would make illegal continuation of their previous trade practices more attractive. Even within the Mesopotamian borderland itself, the enforcement of a restriction on trade to Nisibis alone would have ruined the Euphrates and desert routes. 646 While the fate of the latter route is unclear after the third-century destruction of Palmyra, the Euphrates route certainly seemed to remain in operation. Ammianus’ description of Julian’s expedition shows a number of populous cities in the Euphrates which must have been connected by trade routes. Ammianus specifically notes trade wealth (commercandi opimitate) in his description of Callinicum on the Balikh. If all cross-border trade was restricted to Nisibis and the northern route, then exchange along the Euphrates route would have been severely curtailed. 647 For many , the savings to be gained by non-compliance would have outweighed the risks. If all trade and movement was channelled through one point, why was Nisibis chosen as that point? The choice of Nisibis reflects the long-standing regional prominence of the city and indicates the importance of the northern route in the fourth century . Following the fall of Hatra in 240 CE, trade from Babylonia, across the eastern Jazira through Singara to the cities at the foot of the Tur Abdin seems to have 646 For these routes, see Chapter 5. 647 Young (2001) 188–90; Chaumont (1984) 185/448 shifted to the Tigris. 648 The Tigris was only navigable as far north as Bezabde, so disembarking and continuing inland through Nisibis was a natural option which made as much use of riverine transportation as possible. 649 Nisibis was a large city with a fertile hinterland at the eastern end of the northern routes to Syria and western Anatolia and with access to routes north into Armenia. 650 Although it lay around 100 km west of the Tigris, not on the Tigris as Petrus records (Νισιβιν την πολιν παρακειμενην τῷ Τιγριδι), it was the closest major city on the route to that river. The importance of the city in the Roman urban network of the borderland is indicated by its status as a colonia. Three of the other coloniae (Carrhae, Edessa, Reshaina) of Osrhoena and Mesopotamia lay further west on the route, while the last, Singara, was much smaller, held less fertile territory , and seems to have had a primarily military role which included supervision over the pastoralists around the Jebel Singar. 651 Nisibis itself was a strategically important city . It occupied a strategic position in a small fertile dry- farming zone between the arid plains to the south and Tur Abdin to the north. It first appears in Roman history resisting an attack by Lucullus with its impressive defences. 652 Festus reports that the city was besieged unsuccessfully three times during Constantius’ reign (337-361). 653 When Jovian surrendered the city in the peace of 363, Ammianus, lamenting the act, calls Nisibis the strongest bulwark of the Orient (orientis firmissimum claustrum). 654 It served as the capital of Mesopotamia provincia, headquarters of the dux 648 Dignas (2007) 196. There is evidence of demographic shifts from the region of Hatra towards the Tigris, Hauser (2000). 649 Dignas (2007) 197. 650 For the fertility of Mygdonia, see Chapter 1; for the physical geography and urban history of Nisibis, see also Appendix 1.13: Nisibis; for the northern route across Mesopotamia, see Chapter 5. 651 Millar (1993) 144. Singara: Amm. Marc. 20.6.8-9; Kennedy and Riley (1990) 125–131, figs. 73–75; Appendix 1.15: Singara. 652 Dio 36.6. 653 Festus Brev. 27.2. 654 Amm. Marc. 25.8.14: “Constabat enim orbem eoum in dicionem potuisse transire Persidis, nisi haec civitas habili situ et moenium magnitudine restitisset. Miseri tamen licet maiore venturi pavore constringerentur, spe tamen sustentari potuerunt exigua, hac scilicet, quod velut suopte motu vel exoratus eorum precibus imperator eodem statu retinebit urbem, orientis firmissimum claustrum.” (For it was clear that the entire Orient might have passed into the control of Persia, had not this city with its advantageous situation and mighty walls resisted him. Nevertheless, however much the unhappy people were tormented with great fear of the future, yet they could 186/448 Mesopotamiae and a major mustering point of the magister militum per Orientem. 655 The important strategic role of Nisibis on the northern route can be seen in its replacement by Dara in the sixth century . After Rome accepted that it had lost Nisibis for good, Anastasius founded and fortified Dara on the northern route just the west of Nisibis. It was immediately a source of tension, although armed conflict did not break out until later in the sixth century. 656 The same considerations applied in 363 CE. Having been acclaimed as emperor by the army following Julian’s death on campaign in Persian territory and needing to secure his rule against internal contenders, Jovian made a treaty with Shapur II to allow him and his army to return to Mesopotamia. 657 Ammianus records the terms: Petebat autem rex obstinatius, ut ipse aiebat, sua dudum a Maximiano erepta, ut docebat autem negotium pro redemptione nostra quinque regiones Transtigritanas: Arzanenam et Moxoenam et Zabdicenam itidemque Rehimenam et Corduenam cum castellis quindecim et Nisibin et Singaram et Castra Maurorum, munimentum perquam oportunum. 658 The terms were the surrender of the Transtigritanian provinces. Ammianus names five regions, but the names differ from those recorded by Petrus Patricus: Arzanene, Zabdicene, and Corduene appear in both lists, but Ammianus names Rehimene and Moxoene, instead of Intelene and Sophene. This may indicate that the relative importance of these kingdoms between the Tigris and the Taurus had changed over the intervening 65 years. More importantly however, was the surrender of 18 fortified sites along the frontier: 15 unnamed forts (castellis quindecim) and well as Nisibis, Singara and Castra Maurorum (the Moors’ Fort). Ammianus sustain themselves with one slight hope, namely , that the emperor would, of his own accord or prevailed upon by their entreaties, keep the city in its present condition, as the strongest bulwark of the Orient.) Loeb trans. Lightfoot (1988) 105–6; Dignas (2007) 199–200. 655 Amm. Marc. 14.9.1; 19.9.6 and 20.6.9. Lightfoot (1988) 107. 656 Dignas (2007) 100–104; Greatrex and Lieu (2002) 74–77. 657 Ball (2000) 25–26; Dignas (2007) 133. 658 Amm. Marc. 25.7.9: “However the king stubbornly sought as our ransom, the five Transtigritanian regions which were his and which Maximianus had taken from him, as he claimed, a short while before as negotiations showed: Arzanena, Moxoëna, and Zabdicena, as well as Rehimena and Corduena with fifteen fortresses and Nisibis, Singara and Castra Maurorum, an exceedingly important stronghold.” 187/448 reckons this as an overly high cost. 659 The sites were to be turned over empty of their populations, sparking what must have been a considerable movement of soldiers and civilians from the border zone towards the centre of the empire. 660 There is no notice of any other changes. As far as we know , Nisibis remained the “place for transactions” (τοπον τῶν συναλλαγμάτων) as stipulated in the treaty of 298. Although Ammianus is only concerned with the military and moral effects of the treaty , later writers indicate that Nisibis retained this important role as a focus of inter-imperial commercial activity . 661 The burden of decentralised control of movement and tax revenue switched to Rome while the economies of scale still lay at Nisibis, now in Persian hands. This centralisation ended with an edict of 408/9 which prohibited merchants from going further than Nisibis, Callinicum and Artaxata to buy and sell goods. 662 As mentioned above, the stated aim of this edict 659 Amm. Marc. 25.7.10: “Et cum pugnari deciens expediret, ne horum quicquam dederetur...” (And while it would have been better to fight ten battles than give up any of these...) He characterised Jovians surrender of the sites and relocation of the loyal Roman subjects as a betrayal: Amm. Marc. 25.9.8 “...Nisibi prodita, quae iam inde a Mithridatici regni temporibus, ne oriens a Persis occuparetur...” (...betraying Nisibis, which ever since the time of King Mithridates' reign had resisted with all its might the occupation of the Orient by the Persians.) Loeb trans. 660 Amm. Marc. 25.7.11: “hac perniciosa verborum ille adsiduitate nimia succensus, sine cunctatione tradidit omnia quae petebantur, difficile hoc adeptus ut Nisibis et Singara sine incolis transirent in iura Persarum, a munimentis vero alienandis reverti ad nostra praesidia Romani permitterentur.” (Jovian, inflamed by these dangerous hints too continually repeated, without delay surrendered all that was asked, except that with difficulty he succeeded in bringing it about that Nisibis and Singara should pass into control of the Persians without their inhabitants, and that the Romans in the fortresses that were to be taken from us should be allowed to return to our protection.) Loeb trans. 661 Eutropius 10.17; Festus 29; Ios. Styl. 18: complaint that revenues of Nisibis were so high. 662 CJ 4.63.4: “Imperatores Honorius, Theodosius. Mercatores tam imperio nostro quam persarum regi subiectos ultra ea loca, in quibus foederis tempore cum memorata natione nobis convenit, nundinas exercere minime oportet, ne alieni regni, quod non convenit, scrutentur arcana. 1. Nullus igitur posthac imperio nostro subiectus ultra Nisibin Callinicum et Artaxata emendi sive vendendi species causa proficisci audeat nec praeter memoratas civitates cum persa merces existimet commutandas: sciente utroque qui contrahit et species, quae praeter haec loca fuerint venumdatae vel comparatae, sacro aerario nostro vindicandas et praeter earum ac pretii amissionem, quod fuerit numeratum vel commutatum, exilii se poenae sempiternae subdendum.” (The same Emperors [Emperors Honorius and Theodosius] to Anthemius, Praetorian Prefect. Merchants subject to our sway or that of the kind of Persia, must in no event hold markets in places that lie beyond those that were agreed on in the treaty with that nation, so that they may not improperly spy into the secrets of another kingdom. 1. No one, therefore, subject to our rule, shall thereafter dare to go beyond Nisibis, Callinicum and Artaxata for the purpose of buying or selling any wares, nor think of exchanging wares with a Persian outside of the cities mentioned; 188/448 was to prevent the movement of sensitive information across the border. The three sites to which merchants could travel were along the northern and Euphrates routes and in Armenia respectively . If Nisibis was the sole site of trade, then this reflects a reopening of trade routes, especially along the Euphrates which had been so important in earlier periods. It is possible that this route never went out of operation and that either Petrus Patricus and the Expositio simply focused on Nisibis and excluded Callinicum, or that the trade still flowed along the Euphrates route (and through Armenia) but with some kind of reporting function at Nisibis. Ammianus’s reference to the Transtigritanian provinces brings their role in the Mesopotamian borderland to our attention. Petrus Patricus reported that Rome was to have control of Intelene, Sophene, Arzanene, Karduene and Zabdikene, five regions between the Taurus mountains and the banks of the upper Tigris. 663 These seem to have been in the possession of the Armenian king, so they were not strictly territorial concessions on the part of the Persians. Direct or indirect control of the Armenian throne was a major point of dispute between the Roman and Iranian empires throughout their period of contact. Control of those Transtigritanian provinces meant control of the passes between Armenia and Mesopotamia. Because they were situated on the southern slopes of the Taurus, at least two of those provinces were included by Strabo in Mesopotamia: Arzanene was the site of Tigranocerta and Karduene was part of Strabo’s Gordyene. 664 The name Zabdikene was probably related to the fortress Bezabde (Pinaka) which Strabo places in Gordyene. 665 What exactly it meant for the Romans to have (ἔχειν) these provinces is unclear. The treaty also and both contracting parties may know that the property which is sold or bought outside of these places will be confiscated for the imperial treasury , and in addition to the loss thereof, and the loss of the price which was paid or exchanged, the guilty parties will suffer perpetual exile.) Blume trans. 663 Festus includes a note on the treaty in his Breviarium. He writes (14.25) that that the frontier above the Tigris was reformed (supra ripam Tigridis limes est reformatus) and that the Romans held dominion over (ditionem adsequeremur) five gentes beyond the Tigris. Dillemann (1962) 217–18. 664 Strabo 16.1.23-24; see Chapter 1. For the location of Tigranocerta in Arzanene, see Appendix 1.18: Tigranocerta. 665 Amm. Marc. 20.7.1; Strabo 16.1.24. 189/448 specified that the border was to be the Tigris (τὸν Τιγριν ποταμὸν ἑκατερας πολιτειας ὁροθεσιον ειναι), yet these provinces were on the far side of that river as their collective name (regiones Transtigritanas.) suggests. In Chapter 1 I argued that Strabo’s conception of the Tigris actually begins with the Botan which arises in Lake Van and flows into the Tigris proper. Even if this were the case here, at least Korduene and Zabdikene lie east of that river as well. The use of the Tigris as an inter-imperial political boundary must only apply south of Cizre, once the river enters the alluvial plains of Mesopotamia themselves. 666 Dignas and Winter suggest that this treaty made the Armenian rulers of the Transtigritanian provinces responsible to Rome, but not under direct Roman control. 667 This may have placed them in a similar relationship with Rome as earlier client states like Commagene and Osrhoene. However, Lucius Verus’ Mesopotamian garrisons provides a different model of Roman influence in the region. Rome may have established garrisons at key points in the Transtigritanian provinces intended to control particular strategic points, like the Taurus passes into Armenia or the Armenian nobles who ruled the region. A treaty with the Sasanid king would have legitimised such a deployment. Regardless of how this clause was implemented in the region, it reflects the presence of network connections between Romans and Armenians on at least a military and political level. Palmyra as a borderland city The final polity on which I focus is Palmyra. Whereas the previous treatments took a longer chronological view in order to examine changes in the geographical representation of the space through historical contexts, this section takes Pliny’s description of the site as its focus. In fact, this is our only 666 The Tigris appears as a military and administrative boundary between the two empires in Ammianus. At 18.5.3, a Roman buys property at Hiaspis on the banks of the river so that he may later cross undetected to defect to the Persians; the area is described as the furthest limit of the Roman empire (ad extremas Romani limitis partes). The location of Hiaspis is unknown. When the Persians invaded Roman Mesopotamia near Nisibis in 359 CE, the Tigris is the border through which the Persians burst forth (erupisse hostium vastatorias manus superato flumine, Amm. Marc. 18.6.8-9.) 667 Dignas (2007) 126–28; cf Jones (1971) 223–24. 190/448 geographical description of Palmyra. Pliny places Palmyra in an ambiguous political position between the Parthian and Roman empires, although the chronological period which his description reflects is unclear. A later report by Appian purports to describe conditions in the first century BCE, but seems more indicative of his own time. As a coda to this section, I briefly sum the third century rise and fall of Palmyra. The political and legal position of Palmyra relative to Rome has been much debated. In his section on Roman Syria, Pliny the Elder described the city’s location physically , politically and mathematically: 668 Palmyra urbs nobilis situ, divitiis soli et aquis amoenis, vasto undique ambitu harenis includit agros, ac velut terris exempta a rerum natura, privata sorte inter duo imperia summa Romanorum Parthorumque, et prima in discordia semper utrimque cura. Abest ab Seleucia Parthorum, quae vocatur Ad Tigrim, CCCXXXVII [m.]p., a proximo vero Syriae litore CCIII et a Damasco XXVII propius. 669 The final clause of the first sentence (et prima in discordia semper utrimque cura) is usually taken as an inaccurate (or at least uncorroborated) statement of Palmyra’s military importance between the two empires. This problem arises from an assumption that the final clause implies that both empires moved to secure it militarily at the first sign of conflict, which the sources do not support. However to describe Palmyra as a “cura” need not mean that it was the subject of unattested military campaigns or manoeuvres. The precise legal position of Palmyra within the empire remains somewhat unclear: was the city a provincial centre subject to the governor of Syria, or an independent city-state with some degree of monitoring by the Roman state. 670 The reality seems to lie between the two. Palmyra was incorporated into Roman Syria between 12 and 17 CE, prior to Germanicus’ visit to the city in 17; 671 however, Palmyra may have been the only 668 For a discussion of the physical description of Palmyra, see Chapter 1. For Palmyra’s position in inter-imperial and local trade networks, see Chapter 5. 669 Pliny , NH 5.88: “Palmyra is a city famous for its position, the richness of its soil, and the quality of its water, its fields surrounded on all sides by a vast circuit of sand, as if cut off from the world by nature itself, a private lot between two great empires of Rome and Parthia, and at the first sign of discord between them, always a concern to both. It is distant 337 miles from Seleucia of the Parthians, generally known as Seleucia on the Tigris, 203 from the nearest part of the Syrian coast, and twenty-seven less from Damascus.” 670 Will (1985) etc. : Elton (1996) 90–92. Gawlikowski (1994), 671 Seyrig (1932) 266–77. 191/448 provincial polis to maintain a standing military force. In the Julio-Claudian period, Palmyra developed into an important centre for and facilitator of trans-Arabian trade. 672 Palmyrene forces escorted caravans across the desert and served as garrisons in a number of important caravan termini across the region, including Dura Europus, Hit, and other cities on the Parthian parts of the Euphrates. Regardless of its status in Pliny’s time, Palmyra maintained a standing military force long after most client kingdoms in the region had been directly incorporated into the Roman provincial system. 673 No doubt the loyalty of this military force was a concern to the governors of Syria. It certainly was after the third century when a Roman camp was built nearby following the defeat and capture of Zenobia. 674 Palmyra’s position between the two empires gave her a unique position in the borderland, one that Pliny highlights by linking Palmyra to the Parthian capital Seleucia ad Tigris (Seleucia-Ctesiphon), the Mediterranean and Roman Damascus. In the ancient world, information moved at the speed of travellers. The people moving along the Palmyrene trading network could be a valuable source of strategic military intelligence in the event of political tension. 675 This could be a explanation of the kind of cura these two empires were concerned about. The distances Pliny gives between Palmyra and the sites in its network were derived from the movement of travellers and were not particularly accurate. 676 However, it is important to note that for Pliny’s purposes, their absolute accuracy was less important than what they said about the connections between Palmyra and the two empires. Namely that Palmyra was connected to the heart of the Parthian Empire, Babylonia, and the heart of the Roman empire, the Mediterranean. Large scale, long distance trade relies on personal connections between buyer and seller. These connections often extended beyond commercial 672 See Chapter 5. Palmyra’s association with trading hints and a play on words in Pliny’s description of the city as a privata sorte: sorte can mean principal, in the sense of capital bearing interest. 673 No doubt the loyalty of this military force was a concern to the governors of Syria. It certainly was after the third century when a Roman camp was built nearby following the defeat and capture of Zenobia, Kennedy (1990) 134–7, figs. 82–3. 674 Kennedy (1990) 134–7, figs. 82–3. 675 Lee (1993). 676 See Chapter 5. 192/448 interests. For example, in the sphere of art, Palmyrene sculpture, painting and architecture shows the influence of Greece, Parthia, and the other cities of Syria, melded together in a uniquely local style. 677 Palmyrene network links extended anywhere their trade network could be found, including Dura Europus and other sites along the Euphrates, Hatra, Egypt, Mesene at the head of the Persian Gulf, Mesene’s outpost at Thilouan/Tylos (modern Bahrain), and northwest India. 678 The main problem with Pliny’s description is deciding to when it refers. Because Pliny’s description does not seem to match the political conditions that would hold in a Roman provincial city of the Flavian era, it is widely accepted that it is not an accurate representation of the political conditions of Pliny’s own time. Gawlikowski suggests that Pliny is reporting the situation in the first century BCE at the latest, presumably reflecting the conditions which led Antony to attack the city in 41 BCE. 679 However, as Hekster and Kaizer have argued, those events are reported by Appian in the second century CE and reflect conditions in Appian’s own time. Appian writes: ἀποπλευσάσης δὲ τῆς Κλεοπάτρας ἐς τὰ οἰκεῖα, ὁ Ἀντώνιος ἔπεμπε τοὺς ἱππέας Πάλμυρα πόλιν, οὐ μακρὰν οὖσαν ἀπὸ Εὐφράτου, διαρπάσαι, μικρὰ μὲν ἐπικαλῶν αὐτοῖς, ὅτι Ῥωμαίων καὶ Παρθυαίων ὄντες ἐφόριοι ἐς ἑκατέρους ἐπιδεξίως εἶχον (ἔμποροι γὰρ ὄντες κομίζουσι μὲν ἐκ Περσῶν τὰ Ἰνδικὰ ἢ Ἀράβια, διατίθενται δ’ ἐν τῇ Ῥωμαίων), ἔργῳ δ’ ἐπινοῶν τοὺς ἱππέας περιουσιάσαι. Παλμυρηνῶν δὲ προμαθόντων καὶ τὰ ἀναγκαῖα ἐς τὸ πέραν τοῦ ποταμοῦ μετενεγκάντων τε καὶ ἐπὶ τῆς ὄχθης, εἴ τις ἐπιχειροίη σκευασαμένων τόξοις, πρὸς ἃ πεφύκασιν ἐξαιρέτως, οἱ ἱππέες τὴν πόλιν κενὴν καταλαβόντες ὑπέστρεψαν, οὔτε ἐς χεῖρας ἐλθόντες οὔτε τι λαβόντες. 680 677 Seyrig (1950) 6–7. 678 For sites on the Euphrates, see Smith (2013) 145 and Chapter 5. Hatra: Yon (2013); Egypt: Casson (1989) 20; Tomber (2008) 80; Smith (2013) 161–62. Mesene: Chaumont (1974); Gawlikowski (2007) 129; Young (2001) 144–48; Bahrain: Gawlikowski (2007) 129; India: Smith (2013) 77. 679 Gawlikowski (1994) 27–28; Will (1985); Elton (1996) 90; Gawlikowski (2007). 680 App. BC 5.9: “When Cleopatra had sailed home, Antony sent his horsemen to plunder Palmyra, a polis which was not far from the Euphrates, accusing them of a small matter, that they , being on the boundary between the Romans and the Parthians, interacted with both sides (for as merchants they carry Indian and Arabian goods from the Persians and dispose of them among the Romans), but by this deed he thought to enrich his horsemen. But the Palmyrenes learned about it beforehand and carried their essentials to the other side of the river and onto the bank, preparing themselves with bows with which they are naturally talented, in case anyone should attack them. The horsemen, seizing the empty city , turned around, not having met anyone, not having taken anything.” 193/448 Appian calls the Palmyra of 41 BCE a polis, but also implies that the Palmyrenes were able to avoid the Roman attack by carrying off all their essentials (τὰ ἀναγκαῖα) and leaving the city empty (κενὴν). Archaeological evidence certainly suggests that there was some form of permanent settlement around this time, in particular, a Palmyrene inscription records the erection of a statue in 44 BCE. 681 As Hekster and Kaizer argue, the idea of a wealthy trade city which is able to transport away all of their wealth in the face of a Roman attack is logistically problematic. 682 They argue that the passage contains a number of errors and contemporary resonances, including the present tense of the parenthetical description of Palmyra as a trading town, the apparent location of Palmyra near the Euphrates, and the Parthian flavour of the Palmyrenes in this passage, suggest that Appian is thinking of contemporary affairs for at least part of his account. 683 In particular, we should consider Appian’s statement that “as merchants they carry Indian and Arabian goods from the Persians and dispose of them among the Romans” (ἔμποροι γὰρ ὄντες κομίζουσι μὲν ἐκ Περσῶν τὰ Ἰνδικὰ ἢ Ἀράβια, διατίθενται δ’ ἐν τῇ Ῥωμαίων) as reflecting the conditions of the second century CE rather than the first century BCE. In the dynastic weakness of the third century , Septimius Odainathos, a native Palmyrene whose family had held Roman citizenship since Septimius Severus, rose to power in Palmyra. 684 Following Valerian’s capture, Odainathos, now essentially the ruler of Palmyra, “defended Syria, recovered Mesopotamia” and campaigned as far as Ctesiphon, nominally as a general of Gallienus. 685 In 267/68, Odainathos and his son and heir designate, Hairan, were murdered at Emesa; Odainathos’ wife Zenobia assumed control as regent of 681 Hekster and Kaizer (2004) 72–73; Seyrig et al. (1975) 1.227; Hillers and Cussini (1995) no.1524 (wrongly dated to 44 CE). 682 Hekster (2004) 73–74. 683 Hekster (2004) 78–80. There is also a literary trope at play in the passage, namely a reversal of fortune motif for Antony , Hekster (2004) 75–78. 684 Sartre (2005) 351–3 Sartre suggests that citizenship may have been awarded for services rendered against local adherents of Pescennius Niger in 194. 685 Eutropius 9.10: “Defensa Syria, recepta Mesopotamia usque ad Ctesiphontem Odenathus penetravit”; Dodgeon and Lieu (1991) 71–7 Aurelius Victor (Lib. Caes. 33.3) reports that the Sasanians still controlled Mesopotamia (Dodgeon (1991) 85). 194/448 their other sons. 686 By the end of 270, Zenobia had seized control of the Roman east from Egypt to northern Syria, founded an eponymous fortress on the Euphrates, and had her oldest son Vaballathos take imperial titles. 687 Two years later, Aurelian defeated the Palmyrenes in several battles, seized Antioch, Emesa and Palmyra itself, and took Zenobia to Rome to be led in his triumph. 688 Palmyra survived, but after rumours of an unrest there, a legion was stationed near the city . The subsequent reduction in Palmyrene independence hampered their liminal position as controllers of the desert trade routes between Syria and Babylonia and traffic moved north to the Mesopotamian routes over the following century , reducing the importance of the city . Palmyra’s political position in the Mesopotamian borderland was ambiguous from Rome’s arrival in the Near East and remained so until the third century brought an end to Palmyrene power. Palmyra’s position commercially and geographically between Roman and Parthian space allowed the city to maintain a liminal position in the military , political and legal spheres as well. It seems that Palmyra’s connections to the peoples and states to the east were sufficiently valuable to Rome that a degree of independence in certain other areas were permitted or perhaps even encouraged. The maintenance of a standing military force may have fallen into this category . These various ambiguities culminated and ultimately collapsed with the third century rise of Odainathos and Zenobia. Conclusions The steady expansion of Roman networks of control into local political and social organisations can be seen in the works of the imperial geographic writers. As they proceeded chronologically , so the leading edge of Roman power slowly moved from west to east over the course of the first four centuries CE. Strabo describes Commagene as a wealthy and strategically important part of Roman space. The structure of his 686 Sartre (2005) 355 Dodgeon (1991) 80–4 687 Sartre (2005) 356–7 Dodgeon (1991) 85–9 688 Sartre (2005) 357–8 Dodgeon (1991) 92–107 The sources differ on Zenobia’s ultimate fate. 195/448 narrative emphasises the place of the former kingdom at the edge of Roman space. By contrast, Pliny barely touches on Vespasian’s new annexation. To the extent that it is described, Commagene is not distinguished from the rest of Roman Syria. Just east of the Euphrates, Edessa appears only vaguely in Strabo’s Mesopotamia, as confused amalgam with Hierapolis/Bambyce, and Osrhoene not at all. By Pliny’s time, and in his narrative, Osrhoene occupied a much more important place in the Mesopotamian borderland. He describes the people and its cities in an ambiguous space between the two empires, narratively placed within the sphere of both. The ambiguity of the third century applies to geographical texts even more than historical texts. When the situation becomes clearer in the mid-fourth century , the geographical area of Mesopotamia had become politically and administratively divided between Sasanid space, and the Roman provinces of Osrhoena and Mesopotamia, the former centred on Edessa and Osrhoene, the latter on Nisibis. That city became an focal point for political, military and economic activity as the border region became more rigid. A series of treaties and laws attempted to control the movement of people and information across this increasingly tense and militarised border between the Sasanian and Roman empires. The treatment of Palmyra in the sources is in some ways typical of the borderland polities encountered in the geographical sources. It does not appear consistently in all of them, it is linked to mobility , and it is placed in an ambiguous political position. Palmyra only appears in Pliny , is connected to the Babylonian and Mediterranean worlds by routes of movement, and is presented as sitting in an ambiguous political space between the two empires it connects. The same analogy could be used of the representations of Commagene, Edessa and Osrhoene, and Nisibis. 196/448 Chapter 3: Using Hellenistic Knowledge At the beginning of his description of Mesopotamia in book six of his Natural History , Pliny the Elder sets the tone of this part of the narrative with a focus on the Macedonian impact on Mesopotamian history: Mesopotamia tota Assyriorum fuit, vicatim dispersa praeter Babylona et Ninum. Macedones eam in urbes congregavere propter ubertatem soli. 689 As we saw in his reference to the Assyrian origins of praefectura Mesopotamia, Pliny conceived of the Assyrians as inhabiting a broad area including both upper and lower Mesopotamia and the plains on the east bank of the Tigris. 690 Here Pliny uses Assyrii to describe the local Aramaic-speaking population whom he presents as scattered in villages and largely lacking cities prior to the arrival of the urbanising Macedonians. Seleucid rule favoured the exercise of royal power through the polis structures, but while there were several waves of urban foundation throughout the Hellenistic period, many of these “foundations” simply involved the establishment of Seleucid political structures in existing local settlements. 691 Pliny’s brief and reductive description of the complex political and demographic processes of Mesopotamian settlement history prioritises the Macedonian urbanisation that would be modified and inherited by Roman provincial structures and was valued as a significant indicator of “civilisation” in Roman provincial ideology . 692 The contrast Pliny draws between Mesopotamia Assyriorum and Macedones (NH 6.117) serves to place the role that Macedonian rulers played in the productivity of Mesopotamia squarely in the foreground. 689 Pliny , NH 6.117: “All Mesopotamia belonged to the Assyrians, the population scattered in villages except for Babyon and Ninevah. The Macedonians gathered them together in cities because of the fertility of the soil.” 690 Pliny NH 5.86: “ab Assyriis originem trahens”. For the area which Pliny considered the Assyrians to inhabit, see Chapter 1. 691 Cohen (2006); Jones (1971). 692 On the importance of Hellenistic urbanisation to Roman provincial ideology , see Ando (2012). For pre- Hellenistic settlement history , see Liverani (1988); Fales (1990); Kühne (1995); Wilkinson et al. (1994); Wilkinson (1995); Wilkinson (2000); Wilkinson et al. (2005). 197/448 In highlighting the role of Macedonian colonisation and cultural influence in Mesopotamia, Pliny makes explicit what all the geographical sources display in various forms: the important role that Macedonian colonisation and control had in shaping the region that would become the Mesopotamian borderland. The Hellenistic sources on which the Roman authors relied were partially responsible for this focus. Those writers naturally prioritised the achievements of the culture to which they laid claim, and when the Romans came to consider themselves the heirs to that Hellenistic cultural heritage, they turned to those Hellenistic works to provide not just information about the world as understood by Hellenistic thinkers, but to lend a certain cultural authority to their own work. However, too often the authors of the Roman era are considered little more than imitators of their Hellenistic forebears. The geographical writers working in the context of the ferox victor were far from the passive recipients of their predecessors’ artes. This chapter will show how those geographical writers, particularly Strabo and Pliny , drew on Hellenistic sources of knowledge and fashioned them into a Roman geography of Mesopotamia. The first half of this chapter examines the source methodology of Strabo, Pliny and Ptolemy . I examine their attitudes to personal, first-hand knowledge gathered through autopsy and literary , second-hand knowledge gathered through the study of earlier scholarship. I then investigate who their sources of geographical data were and how they identified and deployed them in their own work. Who and how the geographical writers explicitly cited provides a starting point from which to consider which sources they implicitly used for geographical material but did not name. In this first part, I do not aim at an exhaustive Quellenforschung, but rather to show the intellectual basis of these geographical works and, moreover, the modes and traditions of knowledge to which they appeal and from which they draw their authority . As we shall see, Hellenistic material plays a fundamental role in the construction of these Roman geographies, even when it is not explicitly acknowledged. In the second half of the chapter, I build on these examinations of geographical source methodology to analyse the descriptions of the Mesopotamian borderland. In particular, the role in 198/448 these narratives of Greek and Roman traditions of spatial measurement, systems of delimitation and denomination, and the way geographical material gathered and recorded in the Hellenistic period formed the backbone of these geographical narratives and constituted an appeal to the authority of Hellenistic traditions. Strabo’s sources Strabo was culturally Greek and explicitly placed himself in the Greek geographical tradition. He traced his intellectual lineage back to Homer whom he regarded as the first geographer. 693 Strabo made extensive use of the Hellenistic geographic tradition, especially Eratosthenes, whom he spends much of his introduction criticising, Posidonius, himself a native Syrian from Apamea, and Artemidorus. 694 Strabo’s engagement with his predecessors is such that Romm called the Geography a “treasure-trove of the fragments of Hellenistic critical debate.” 695 Pothecary goes further: “Strabo is so enmeshed in the Greek culture of his day that he cannot struggle free of it: his way of looking at the world is subjectively determined by his immersion in the Greek culture of which he is so proud.” 696 There is a large degree of 693 Dueck (2000) 31–40; Strabo 1.1.2: “καὶ πρῶτον ὅτι ὀρθῶς ὑπειλήφαμεν καὶ ἡμεῖς καὶ οἱ πρὸ ἡμῶν, ὧν ἐστι καὶ Ἵππαρχος, ἀρχηγέτην εἶναι τῆς γεωγραφικῆς ἐμπειρίας Ὅμηρον.” (First, I say that both I and my predecessors, one of whom was Hipparchus himself, are right in regarding Homer as the founder of the science of geography) Loeb trans. Strabo 1.1.11: “Νυνὶ δὲ ὅτι μὲν Ὅμηρος τῆς γεωγραφίας ἦρξεν, ἀρκείτω τὰ λεχθέντα. φανεροὶ δὲ καὶ οἱ ἐπακολουθήσαντες αὐτῷ ἄνδρες ἀξιόλογοι καὶ οἰκεῖοι φιλοσοφίας, ὧν τοὺς πρώτους μεθ’ Ὅμηρον δύο φησὶν Ἐρατοσθένης, Ἀναξίμανδρόν τε Θαλοῦ γεγονότα γνώριμον καὶ πολίτην καὶ Ἑκαταῖον τὸν Μιλήσιον· τὸν μὲν οὖν ἐκδοῦναι πρῶτον γεωγραφικὸν πίνακα, τὸν δὲ Ἑκαταῖον καταλιπεῖν γράμμα, πιστούμενον ἐκείνου εἶναι ἐκ τῆς ἄλλης αὐτοῦ γραφῆς.” (For the moment what I have already said is sufficient, I hope, to show that Homer was the first geographer. And, as every one knows, the successors of Homer in geography were also notable men and familiar with philosophy. Eratosthenes declares that the first two successors of Homer were Anaximander, a pupil and fellow-citizen of Thales, and Hecataeus of Miletus; that Anaximander was the first to publish a geographical map [γεωγραφικὸν πίνακα], and that Hecataeus left behind him a work on geography , a work believed to be his by reason of its similarity to his other writings. ) Loeb trans. Strabo is particularly critical of Eratosthenes’ dismissal of Homer as a useful geographical source, Dueck (2000) 34. 694 Eratosthenes: 276-195 BCE; Posidonius: 135-51 BCE; Artemidorus: 2 nd -1 st C BCE. Strabo names four geographers whom he deems worthy of engagement in philosophical terms: Eratosthenes, Hipparchus, Posidonius and Polybius (Strabo 1.2.1). On Hipparchus, see Dueck (2000) 58–59; Viedebantt (1915); Dicks (1960); Neugebauer (1975) 274–77, 332–38; Hübner, ‘Hipparchus [6]’, BNP . Strabo (3.5.7) defines Artemidorus as a layman in certain matters (ἰδιώτης) but uses him for distances and measurements, Dueck (2000) 59–60. 695 Romm (1992) 197. 696 Pothecary (2005) 6. 199/448 truth to such statements, but they should not obscure the fact that the geography Strabo constructed did not simply reflect a Greek understanding of the world. Strabo was also a Roman subject, perhaps even a citizen, who placed Rome and Roman power at the centre of his work and aimed at practical scholarship of use to contemporary politicians and generals. 697 The boundaries between Strabo’s reflections of Hellenism and of Roman imperial ideology thus bear investigation. This chapter begins with an examination of Strabo’s attitude towards the acquisition and production of geographic knowledge. I will then discuss the specific sources he uses to construct him image of Mesopotamia and how he used them in his geographical narrative. These investigations will allow some conclusions to be drawn about Strabo’s geographical methods. In book 2, Strabo discussed the acquisition of his geographical material. He begins by stating his credentials as a widely-travelled scholar. ἐροῦμεν δὴ τὴν μὲν ἐπελθόντες αὐτοὶ τῆς γῆς καὶ θαλάττης, περὶ ἧς δὲ πιστεύσαντες τοῖς εἰποῦσιν ἢ γράψασιν. Ἐπήλθομεν δὲ ἐπὶ δύσιν μὲν ἀπὸ τῆς Ἀρμενίας μέχρι τῶν κατὰ Σαρδόνα τόπων τῆς Τυρρηνίας, ἐπὶ μεσημβρίαν δὲ ἀπὸ τοῦ Εὐξείνου μέχρι τῶν τῆς Αἰθιοπίας ὅρων·οὐδὲ τῶν ἄλλων δὲ οὐδὲ εἷς ἂν εὑρεθείη τῶν γεωγραφησάντων πολύ τι ἡμῶν μᾶλλον ἐπεληλυθὼς τῶν λεχθέντων διαστημάτων, ἀλλ’ οἱ πλεονάσαντες περὶ τὰ δυσμικὰ μέρη τῶν πρὸς ταῖς ἀνατολαῖς οὐ τοσοῦτον ἥψαντο, οἱ δὲ περὶ τἀναντία τῶν ἑσπερίων ὑστέρησαν· ὁμοίως δ’ ἔχει καὶ περὶ τῶν πρὸς νότον καὶ τὰς ἄρκτους. 698 Strabo divides his sources of knowledge into two categories: that gathered by personal observation (ἐπελθόντες αὐτοὶ) and that accepted through trust (πιστεύσαντες). This latter category , he divides further 697 The centrality of Roman power is evident throughout the books which describe areas controlled by Rome. Nicolet characterises Strabo as “an Asian Greek, a representative of those intellectual notables who were thoroughly Hellenized and completely won over by the new Roman power, and who—better yet—rubbed shoulders with the governing class and especially with those entrusted with the administration of the vast empire, indeed, with its expansion”, (1991) 73. See also, Dueck (2000) 107–122. Strabo declares his practical purpose at 1.1.22-23 and throughout his work. On the question of Strabo’s citizenship, for which there is no real evidence except for the apparent Latinity of his name, see Pothecary (1999); Dueck (2000) 7–8. 698 Strabo 2.5.11: “Now I shall relate those parts of the land and sea I have visited myself and those parts where I have placed my trust in the words and writing of others. I have travelled towards the west from Armenia as far as the regions of Tyrrhenia opposite Sardinia, and towards the south from the Black Sea as far as the frontiers of Ethiopia. And you could not find anyone else among the geographers who has travelled over much more of the distances just mentioned than I; indeed, those exceeding me in the western part have not reached so much in the east, and those exceeding me in the eastern part, are behind in the western; and the same holds true concerning those parts towards the south and north.” 200/448 between oral and written testimony (τοῖς εἰποῦσιν ἢ γράψασιν), perhaps granting equal weight to both. For geographical knowledge, the first category , personal observation, required travel to observe the various parts of the world to be described. Accordingly , Strabo provides his credentials in this regard. He gives the extent of his travels in terms of the furthest region in each cardinal direction which he has visited: Armenia, Tyrrhenia (Etruria, Tuscany), the Black Sea (τοῦ Εὐξείνου), the frontiers (ὅρων) of Ethiopia. References to autopsy within Strabo’s text fall within these bounds. There has been much scholarly speculation on how widely Strabo travelled and observed the places of which he wrote. 699 His travels in the eastern Mediterranean seem to have been extensive, but there is no solid evidence that he visited Syria and none at all that he personally visited Mesopotamia or any place to the east of the Roman empire. Strabo gives general bounds to his travel here, and alludes to personal visits to places within his text, but he did not consider it important to list or explicitly catalogue his travels. It was enough to claim a superior breadth of travel experience than any previous geographer. Having begun with a claim to a superior degree of personal geographical experience, Strabo proceeds to justify his use of second hand material by an appeal to the science of perception. The idea of personal travel to and autopsy of the far-flung spaces of the word prioritises the sense of sight. Strabo acknowledges the value of travel and the visual sense, but argues for a more holistic approach to sense perception by which the mind collates the input of all of the senses, and particularly relies on hearing. τὸ μέντοι πλέον κἀκεῖνοι καὶ ἡμεῖς ἀκοῇ παραλαβόντες συντίθεμεν καὶ τὸ σχῆμα καὶ τὸ μέγεθος καὶ τὴν ἄλλην φύσιν ὁποία καὶ ὁπόση, τὸν αὐτὸν τρόπον ὅνπερ ἡ διάνοια ἐκ τῶν αἰσθητῶν συντίθησι τὰ νοητά· σχῆμα γὰρ καὶ χρόαν καὶ μέγεθος μήλου καὶ ὀδμὴν καὶ ἁφὴν καὶ χυμὸν ἀπαγγέλλουσιν αἱ αἰσθήσεις, ἐκ δὲ τούτων συντίθησιν ἡ διάνοια τὴν τοῦ μήλου νόησιν· καὶ αὐτῶν δὲ τῶν μεγάλων σχημάτων τὰ μέρη μὲν αἴσθησις ὁρᾷ, τὸ δ’ ὅλον ἐκ τῶν ὁραθέντων ἡ διάνοια συντίθησιν. οὕτω δὲ καὶ οἱ φιλομαθεῖς ἄνδρες ὥσπερ αἰσθητηρίοις πιστεύσαντες τοῖς 699 Dueck (2000) 15–30. From references in his Geography , he visited the Aegean (10.5.3), Egypt and Aithiopia (2.5.12; 17.1.24), and Cappadocia (12.2.3), travelled to Brundisium by sea and from there to Rome by road (6.3.7), Clarke (1997) 99. Dueck rebuts Pais’ suggestion that Strabo visited Berytus and Ascalon, Dueck (2000) 22, n.52. On Strabo’s accompaniment of Aelius Gallus’ expedition down the Arabian coast of the Red Sea, described by Strabo at 16.4.22-24, see Nicolet (1991) 85, with n.4. 201/448 ἰδοῦσι καὶ πλανηθεῖσιν οὓς ἔτυχε τόπους ἄλλοις κατ’ ἄλλα μέρη τῆς γῆς, συντιθέασιν εἰς ἓν διάγραμμα τὴν τῆς ὅλης οἰκουμένης ὄψιν· ἐπεὶ καὶ οἱ στρατηγοὶ πάντα μὲν αὐτοὶ πράττουσιν, οὐ πανταχοῦ δὲ πάρεισιν, ἀλλὰ πλεῖστα κατορθοῦσι δι’ ἑτέρων, ἀγγέλοις πιστεύοντες καὶ πρὸς τὴν ἀκοὴν διαπέμποντες οἰκείως τὰ προστάγματα. ὁ δ’ ἀξιῶν μόνους εἰδέναι τοὺς ἰδόντας ἀναιρεῖ τὸ τῆς ἀκοῆς κριτήριον, ἥτις πρὸς ἐπιστήμην ὀφθαλμοῦ πολὺ κρείττων ἐστί. 700 Strabo places great weight on hearing as sense of geographical perception, by which he refers to the transmission of prior scholarship through the act of scribal reading. He likens the work of the geographer to that of a general who hears second hand reports from others and, trusting them like his own sensory organs (ὥσπερ αἰσθητηρίοις πιστεύσαντες), is able to bring about victory (κατορθοῦσι δι’ ἑτέρων). In Strabo’s model of geographical knowledge production, the geographer gathers sensory perceptions from trusted sources and the mental assembly (συντίθημι) of those varied perceptions into a conception (νόησις) of the geographical space in question. In support of this argument, Strabo provides a list of peoples and places concerning which his contemporaries (οἱ νῦν) are able to say (λέγειν) better things. 701 While it stands to 700 Strabo 2.5.11: “However, they and I receive the greater part [of our geographical knowledge] by hearing and put together shape and size and other natural characteristics of whatever quality and quantity , indeed, the same way in which the mind assembles intelligible things out of perceptible things. The senses report the shape, skin and size of an apple, as well as its smell, feel and flavour, and the mind assembles the concept of an apple from those things. While the senses perceive the parts of large shapes, the mind assembles the whole out of what is perceived. In this way , those men who love knowledge trust just like sensory organs those who have seen and wandered through whatever places they happened upon in this or that part of the earth, and assemble their perception of the whole world into a single figure. Although generals do everything themselves, they are not present everywhere, but successfully manage most things through other people, trusting in messengers and sending around their orders properly in accordance with the reports they have heard. Those who think that only those who have seen possess knowledge remove the criterion of hearing, which is much more important than sight for knowledge.” 701 Strabo 2.5.12: “Μάλιστα δ’ οἱ νῦν ἄμεινον ἔχοιεν ἄν τι λέγειν περὶ τῶν κατὰ Βρεττανοὺς καὶ Γερμανοὺς καὶ τοὺς περὶ τὸν Ἴστρον τούς τε ἐντὸς καὶ τοὺς ἐκτὸς Γέτας τε καὶ Τυρεγέτας καὶ Βαστάρνας, ἔτι δὲ τοὺς περὶ τὸν Καύκασον, οἷον Ἀλβανοὺς καὶ Ἴβηρας. Ἀπήγγελται δ’ ἡμῖν καὶ ὑπὸ τῶν τὰ Παρθικὰ συγγραψάντων, τῶν περὶ Ἀπολλόδωρον τὸν Ἀρτεμιτηνόν, ἃ πολλῶν ἐκεῖνοι μᾶλλον ἀφώρισαν, τὰ περὶ τὴν Ὑρκανίαν καὶ τὴν Βακτριανήν· τῶν τε Ῥωμαίων καὶ εἰς τὴν εὐδαίμονα Ἀραβίαν ἐμβαλόντων μετὰ στρατιᾶς νεωστί, ἧς ἡγεῖτο ἀνὴρ φίλος ἡμῖν καὶ ἑταῖρος Αἴλιος Γάλλος, καὶ τῶν ἐκ τῆς Ἀλεξανδρείας ἐμπόρων στόλοις ἤδη πλεόντων διὰ τοῦ Νείλου καὶ τοῦ Ἀραβίου κόλπου μέχρι τῆς Ἰνδικῆς, πολὺ μᾶλλον καὶ ταῦτα ἔγνωσται τοῖς νῦν ἢ τοῖς πρὸ ἡμῶν. ὅτε γοῦν Γάλλος ἐπῆρχε τῆς Αἰγύπτου, συνόντες αὐτῷ καὶ συναναβάντες μέχρι Συήνης καὶ τῶν Αἰθιοπικῶν ὅρων ἱστοροῦμεν, ὅτι καὶ ἑκατὸν καὶ εἴκοσι νῆες πλέοιεν ἐκ Μυὸς ὅρμου πρὸς τὴν Ἰνδικήν, πρότερον ἐπὶ τῶν Πτολεμαϊκῶν βασιλέων ὀλίγων παντάπασι θαρρούντων πλεῖν καὶ τὸν Ἰνδικὸν ἐμπορεύεσθαι φόρτον.” (In particular the writers of the present time can give a better account of the Britons, the Germans, the peoples both north and south of the Ister, the Getans, the Tyregetans, the Bastarnians, and, furthermore, the peoples in the regions of the Caucasus, such as the Albanians and the Iberians. Information has been given us also concerning Hyrcania and Bactriana by the writers of Parthian histories (Apollodorus of Artemita and his school), in which they marked off those countries more definitely than many other writers. 202/448 reason that someone able to give a good account of a distant people or place would be good source of heard information (ἀκοῇ), Strabo does not elaborate on why that would be better than observing the objects of study himself (ἰδόντας). Regardless, Strabo thus justifies the expansion of his geographical project beyond the bounds of his personal experience of travel and autopsy . W e now turn to the expert opinions which Strabo trusted to assemble his concept (νόησις) of οἱ Ἀσσύριοι, the space which included the Mesopotamian borderland. Strabo mentions many of his sources by name throughout his work. In book 16, he mentions Eratosthenes, Posidonius, Artemidorus, Nearchus and Orthagoras, Aristobulus, and Polyclitus. 702 The first three of these sources were major geographic writers of the Hellenistic period, while the latter four all had first-hand knowledge of Alexander’s campaigns. Many of Strabo’s sources for the area of the former Achaemenid Persian empire wrote based on their first-hand knowledge of the spaces in question. Aristobulus of Cassandria was a Macedonian who accompanied Alexander on his campaigns and later wrote a history of those campaigns. 703 His geographic and ethnographic interests made him an important source for Strabo on Babylonia and India. Strabo cites Again, since the Romans have recently invaded Arabia Felix with an army , of which Aelius Gallus, my friend and companion, was the commander, and since the merchants of Alexandria are already sailing with fleets by way of the Nile and of the Arabian Gulf as far as India, these regions also have become far better known to us of today than to our predecessors. At any rate, when Gallus was prefect of Egypt, I accompanied him and ascended the Nile as far as Syene and the frontiers of Ethiopia, and I learned that as many as one hundred and twenty vessels were sailing from Myos Hormos to India, whereas formerly , under the Ptolemies, only a very few ventured to undertake the voyage and to carry on traffic in Indian merchandise.) Loeb trans. For Apollodorus of Artemita: Gärtner, Hans Armin, “Apollodorus [8a]” BNP . D’Hautcourt, Alexis, “Apollodoros of Artemita (779)”, BNJ; see Chapter 4 for a brief discussion and further references. 702 Eratosthenes: Strabo 16.1.12; Posidonius: 16.1.15; Artemidorus: 16.2.33; Nearchus and Orthagoras 16.3.5; Aristobulus 16.1.11; and Polyclitus 16.1.13. 703 Most of the evidence for Aristobulus’ work comes from fragments preserved by Strabo and Arrian’s Anabasis. Pownall, ‘Aristoboulos of Kassandreia (139)’ BNJ. Arr. Anab. 1.pr.1-2 (= BNJ 139 T6): “Πτολεμαῖος ὁ Λάγου καὶ ᾽Αριστόβουλος ὁ ᾽Αριστοβούλου ὅσα μὲν ταὐτὰ ἄμφω περὶ ᾽Αλεξάνδρου … συνέγραψαν, ταῦτα ἐγὼ ὡς πάντηι ἀληθῆ ἀναγράφω ... ὁ μὲν ὅτι συνεστράτευσε βασιλεῖ ᾽Αλεξάνδρωι ᾽Αριστόβουλος ... ἄμφω δὲ ὅτι τετελευτηκότος ἤδη ᾽Αλεξάνδρου ξυγγράφουσιν ...” (In the places where Ptolemy the son of Lagos and Aristoboulos the son of Aristoboulos reach a consensus about Alexander, I record what they say as entirely true ... in the case of Aristoboulos because he went on campaign with King Alexander ... and because both men wrote their histories after the death of Alexander.) Pownall trans. Brunt (1974); Badian, ‘Aristobulus [7]’, BNP . Aristobulus was still writing after 301 BCE; he reports the death of Antigonus as Ipsus in that year, Arr. Anab. 7.18.5. 203/448 Aristobulus for a long passage on Alexander’s inspection and repair of the canals of Babylonia and the construction of a fleet for his proposed invasion of Arabia. 704 Polyclitus of Larisa also participated in Alexander’s campaigns and wrote a work whose exact nature is unclear from the preserved fragments. 705 Strabo cites Polyclitus to critique the latter’s opinion that the Euphrates never overflows its banks. 706 Nearchus and Orthagoras also participated in Alexander’s campaigns. Nearchus is most notable for his command of the fleet which Alexander dispatched from the mouth of the Indus to survey the coast as far as the mouth of the Euphrates. 707 He later wrote a periplus of this journey which Arrian used extensively for his Indica. Strabo cited this work several times in book 15 which included that coastline within its purview , and twice in book 16 for islands in the Persian Gulf. 708 Orthagoras may have participated in Nearchus’ coastal 704 Strabo 16.1.11: “φησὶ δ᾽ Ἀριστόβουλος τὸν Ἀλέξανδρον αὐτὸν ἀναπλέοντα καὶ κυβερνῶντα τὸ σκάφος ἐπισκοπεῖν καὶ ἀνακαθαίρειν τὰς διώρυγας μετὰ τοῦ πλήθους τῶν συνακολουθησάντων... σκήψασθαι μὲν οὖν αἰτίαν τοῦ πολέμου φησίν, ἐπειδὴ μόνοι τῶν ἁπάντων οὐ πρεσβεύσαιντο οἱ Ἄραβες ὡς αὐτόν, τὸ δ᾽ ἀληθὲς ὀρεγόμενον πάντων εἶναι κύριον...” (Aristobulus says that Alexander himself, when he was sailing up the river, and piloting the boat, inspected the canals, and with his multitude of followers cleared them... Now Alexander alleged as cause of the war, Aristobulus says, that the Arabians were the only people on earth who did not send their ambassadors to him, but in truth was reaching to be lord of all...) Loeb trans. 705 Badian, ‘Polyclitus [4]’ BNP; Sekunda, ‘Polykleitos of Larisa (128)’, BNJ. Polyclitus described the luxury of Alexander and Persian kings, Susa (F2), wonders of India (F9, F10), and waterways of Asia (F5 = Str. 16.1.13; F6 = Str. 15.3.4; F7 – Str. 11.7.4; F11a and b). His work ran to eight books and may have been a history , a collection of anecdotes or a memoir. 706 Strabo 16.1.13 (= BNJ 128 F5): “Πολύκλειτος δέ φησι μὴ πλημμυρεῖν τὸν Εὐφράτην: διὰ γὰρ πεδίων φέρεσθαι μεγάλων, τὰ δ᾽ ὄρη τὰ μὲν δισχιλίους ἀφεστάναι σταδίους τὰ δὲ Κοσσαῖα μόλις χιλίους, οὐ πάνυ ὑψηλὰ οὐδὲ νιφόμενα σφοδρῶς οὐδ᾽ ἀθρόαν ἐπιφέροντα τῇ χιόνι τὴν τῆξιν: εἶναι γὰρ καὶ τὰ ὕψη τῶν ὀρῶν ἐν τοῖς ὑπὲρ Ἐκβατάνων μέρεσι τοῖς προσβορείοις: ἐν δὲ τοῖς πρὸς νότον σχιζόμενα καὶ πλατυνόμενα πολὺ ταπεινοῦσθαι, ἅμα δὲ καὶ τὸ πολὺ τοῦ ὕδατος ἐκδέχεσθαι τὸν Τίγριν καὶ οὕτως πλημμυρεῖν. τὸ μὲν οὖν ὕστατον ῥηθὲν φανερῶς ἄτοπον...” (But Polycleitus says, that the Euphrates does not overflow its banks, because its course is through large plains; that of the mountains (from which it is supplied), some are distant 2000, and the Cossæan mountains scarcely 1000 stadia, that they are not very high, nor covered with snow to a great depth, and therefore do not occasion the snow to melt in great masses, for the most elevated mountains are in the northern parts above Ecbatana; towards the south they are divided, spread out, and are much lower; the Tigris also receives the greater part of the water [which comes down from them], and thus overflows its banks. The last assertion is evidently absurd...) Sekunda trans. 707 Badian, ‘Nearchus [2]’, BNP; Whitby , ‘Nearchos (133)’, BNJ. BNJ 133 T8a = Arr. Ind. 20.1; BNJ 133 T8b = Strabo 15.2.4; BNJ 133 T8c = Diod. Sic. 17.104.3. Curt. 9.10.3; Plut. Alex . 66; Arrian, Anab. 6.19.5; 21.3; Itin. Alex. 51. Badian (1975) provides a critical account of Nearchus’ work. 708 Strabo 16.3.5 (= BNJ 133 F27): εἰρήκασι καὶ Νέαρχος καὶ Ὀρθαγόρας νῆσον... κεῖσθαι... (“Nearchus and Orthagoras have said that an island... lies...”). The island in question is either Tyrian (τυρίνην) or named Ogyros (Ὤγυριν); the former is the manuscript reading, the latter is a common emendation; see Whitby’s commentary on BNJ 133 F27. Strabo 16.3.7 (= BNJ 133 F28): “φησὶ δ᾽ ὁ Νέαρχος τὸν Μιθρωπάστην ἐντυχεῖν...” (Nearchus says that Mithropastes met...) 204/448 expedition. He later wrote an Indoi logoi (Account of India or perhaps Stories of India) of which five fragments have survived. 709 The relative dates of these works are unknown. Strabo cites Orthagoras once in book 16. 710 The other three writers which Strabo cites by name in book 16 were important Hellenistic scholars with excellent access to such first hand accounts, but who had probably not travelled extensively in the regions they described. Eratosthenes of Cyrene was already an established scholar when he was invited to be the head of the library at Alexandria by Ptolemy III Euergetes (246-222 BCE). 711 This position gave him unparalleled access to the scientific and geographical literature of the Hellenistic world, including the ephemerides (field diaries) of Alexander’s campaign. 712 His works included a three volume Geography (the first two books of which concerned the history of the subject and several topical investigations while the third discussed a map of the world) and a volume Περὶ ἀναμετρήσεως τῆς γῆς (On the Measurement of the Earth). 713 Strabo cites and criticises these works extensively , but they were not among those listed in the biography contained in the Suda. 714 Strabo’s criticism of Eratosthenes did not diminish his respect for his predecessor. Rather, it indicated 709 Karttunen, ‘Orthagoras [2]’, BNP; Engels, ‘Orthahoras (713)’, BNJ.; Karttunen (1997) 45–46. The name of Orthagoras’ work is found in Aelianos, Peri zoon idiotetos 16.35 (= BNJ 713 F 3). Orthagoras is usually cited together with Nearchus (BNJ 713 F 1 = BNJ 133 F 12; BNJ 713 F 5 = BNJ 133 F 27) or with Onesicritus (BNJ 713 F 4 = BNJ 134 F 31). Nearchus and Onesicritus are reported together at BNJ 133 F 30-31. Onesicritus was Alexander’s helmsman and interpreter and wrote a biography of Alexander; Goulet-Cazé, ‘Onesicritus’ BNP . Because these three authors are frequently cited together for descriptions of this region, the boundaries between their works is be unclear, see commentary on BNJ 133 F 27 (Whitby ‘Nearchus’). 710 Strabo 16.3.7. Because Strabo only cites Orthagoras once, and in conjunction with Nearchus, Whitby suggests that Strabo may not have read Orthagoras’ account himself, but may have found his opinions in the works of the writers with whom he is cited. He finds support for this suggestion in the use of the singular φημί in 16.3.7 where Strabo is nominally citing both Nearchus and Orthagoras. He takes Νέαρχος alone as its singular subject because other fragments show that Nearchus wrote about the subjects discussed in this passage (See commentary on BNJ 133 F 27). The chronological priority of the two authors’ respective works is not known. 711 Suda, s.v . ᾽Ερατοσθένης = BNJ 241 T 1; Strabo 17.3.22 = BNJ 241 T 2. See commentary on both passages, and Biographical Essay in Pownall, ‘Eratosthenes (241)’ BNJ. Fraser (1970) 181ff.; Blomqvist (1992) 54–55; Nicolet (1991) 61–62. The most recent edition and commentary on the fragments of Eratosthenes is Roller (2010). 712 Blomqvist (1992) 55–56, 58. 713 Zaminer and Tosi, ‘Eratosthenes [2]’ BNP; Fraser (1970) 187–94. 714 BNJ 241 T 1. In fact, were it not for Strabo, we may not even know about Eratosthenes geographic works. 205/448 Strabo’s own desire to define his work in comparison to the great Alexandrian scholar and situate himself in Eratosthenes’ intellectual company . 715 Posidonius (135-51 BCE) was a prolific scholar of the first century CE. 716 He was originally from Apamea, but he became a citizen of Rhodes later in life. 717 He travelled widely and communicated with the elite of Roman society , including the likes of Cicero, Pompey and the Marcelli. 718 Posidonius wrote on many philosophical and scientific subjects and, like Strabo, was both a Stoic and a continuator of Polybius. 719 Strabo shared these intellectual interests with Posidonius and it is no surprise that he included him in his list of predecessors worthy of attention and criticism. 720 715 Strabo 17.3.22 (= BNJ 241 T 2): Κυρηναῖος δ᾽ ἐστὶ καὶ Καλλίμαχος καὶ ᾽Ερατοσθένης, ἀμφότεροι τετιμημένοι παρὰ τοῖς Αἰγυπτίων βασιλεῦσιν, ὁ μὲν ποιητὴς ἅμα καὶ περὶ γραμματικὴν ἐσπουδακώς, ὁ δὲ καὶ ταῦτα καὶ περὶ φιλοσοφίαν καὶ τὰ μαθήματα, εἴ τις ἄλλος, διαφέρων.” (Kallimachos and Eratosthenes are from Cyrene, both honored by the kings of the Egyptians. The former was a poet as well as a scholar of literary criticism, and the latter excelled in these fields and also in philosophy and mathematics, if anyone ever did.) Translation by Pownall, ‘Eratosthenes of Cyrene (241)’ BNJ. Strabo 1.2.1: “εἰ δ᾽ ἀναγκασθησόμεθά που τοῖς αὐτοῖς ἀντιλέγειν, οἷς μάλιστα ἐπακολουθοῦμεν κατ᾽ ἄλλα, δεῖ συγγνώμην ἔχειν: οὐ γὰρ πρόκειται πρὸς ἅπαντας ἀντιλέγειν, ἀλλὰ τοὺς μὲν πολλοὺς ἐᾶν, οἷς μηδὲ ἀκολουθεῖν ἄξιον, ἐκείνους δὲ διαιτᾶν, οὓς ἐν τοῖς πλείστοις κατωρθωκότας ἴσμεν, ἐπεὶ οὐδὲ πρὸς ἅπαντας φιλοσοφεῖν ἄξιον, πρὸς Ἐρατοσθένη δὲ καὶ Ἵππαρχον καὶ Ποσειδώνιον καὶ Πολύβιον καὶ ἄλλους τοιούτους καλόν.” (And if I shall on occasion, be compelled to contradict the very men whom in all other respects I follow most closely , I beg to be pardoned; for it is not my purpose to contradict every individual geographer, but rather to leave the most of them out of consideration – men whose arguments it is unseemly even to follow – and to pass upon the opinion those men whom we recognise to have been correct in most cases. Indeed, to engage in philosophical discussion with everybody is unseemly , but it is honourable to do so with Eratosthenes, Hipparchus, Posidonius, Polybius, and others of their type.) Loeb trans. On this passage and Strabo’s attitude towards his predecessors generally , see Dueck (2000) 56–58; Nicolet (1991) 57. 716 Inwood, ‘Poseidonius [3]’, BNP; Dowden, ‘Poseidonios (87)’ BNJ; Dueck (2000) 60–61; Nicolet (1991) 65. The most recent major edition of the fragments with extensive commentary is Kidd and Edelstein (1989) (= Ed.- Kidd). 717 Strabo 14.2.13 (= BNJ 87 T 2); Strabo 16.2.10 (= BNJ 87 T 3); Ps-Lucian Macrobii 20 (= BNJ 87 T 4); Strabo 7.5.8 (= BNJ 87 T 6); Athenaios, Deipnosophists 6.61 (= BNJ 87 F 4). 718 Travel: Cic. Tusc. Disp. 5.107 = BNJ 87 T 3a. Strabo mentions his travels at various points (3.1.5; 3.2.5; 17.3.4). Circles: Plut. Mar. 45 = BNJ 87 T 7; Strabo 11.1.6 = BNJ 87 T 8a; Plut. Pomp. 42 = T 8b; Pliny NH 7.112 = BNJ 87 T 8c; Cic. Tusc. Disp. 2.61 = BNJ 87 T 8d; See commentary of BNJ 87 T 8a for discussion; Momigliano (1975) 32. 719 Dowden, ‘Poseidonios (87)’ BNJ 87 T1. On Strabo’s Stoicism, see Dueck (2000) 62–69. Strabo (2.2.1) characterised Poseidonius’s geographic writing (in περὶ ὠκεανοῦ) as partly “proper” (τὰ μὲν οἰκείως) and partly mathematical (τὰ δὲ μαθηματικώτερον). 720 Strabo 1.2.1. At 3.2.9, Strabo specifically criticises Posidonius’ rhetorical presentation and emotional engagement with is subject material. Inwood (‘Poseidonius [3]’, BNP) suggests that Posidonius’ moralising presentation of mining (Strabo 3.2.9 = BNJ 87 F 47) was “probably typical of his integrated approach” of presenting natural history combined with philosophical ideas. 206/448 The final source mentioned by Strabo in book 16 is Artemidorus of Ephesus, a geographer of the late second and early first century BCE who wrote a historical work on Ionia and an eleven book Geography (Γεωγραφούμενα) which survives only in fragments, many from Strabo. 721 Artemidorus’ geographic work seems to have taken the form of a periplus; an epitome by Marcianus of Heracleia survives. The only such fragment in book 16 concerns a set of distances between places on the Phoenician coast. 722 Strabo’s first citation of Eratosthenes in book 16, a short passage on the hydrology of the Euphrates, is one of the most complete examples of Strabo’s citation style. ᾿Ερατοσθένης δὲ τῶν λιμνῶν μνησθεὶς τῶν πρὸς τῇ Ἀραβίᾳ, φησὶ τὸ ὕδωρ ἀπορούμενον διεξόδων ἀνοῖξαι πόρους ὑπὸ γῆς καὶ δι᾽ ἐκείνων ὑποφέρεσθαι μέχρι Κοιλοσύρων: ἀναθλίβεσθαι δὲ εἰς τοὺς περὶ Ῥινοκόρουρα καὶ τὸ Κάσιον ὄρος τόπους καὶ ποιεῖν τὰς ἐκεῖ λίμνας καὶ τὰ βάραθρα: οὐκ οἶδα δ᾽ εἰ πιθανῶς εἴρηκεν. αἱ γὰρ τοῦ Εὐφράτου παρεκχύσεις αἱ ποιοῦσαι τὰς πρὸς τῇ Ἀραβίᾳ λίμνας καὶ τὰ ἕλη πλησίον εἰσὶ τῆς κατὰ Πέρσας θαλάττης, ὁ δὲ διείργων ἰσθμὸς οὔτε πολύς ἐστιν οὔτε πετρώδης, ὥστε ταύτῃ μᾶλλον εἰκὸς ἦν βιάσασθαι τὸ ὕδωρ εἰς τὴν θάλατταν, εἴτ᾽ ὑπὸ γῆς εἰτ᾽ ἐπιπολῆς, ἢ πλείους τῶν ἑξακισχιλίων σταδίων διανύειν, ἄνυδρον καὶ ξηρὰν οὕτω, καὶ ταῦτα ὀρῶν ἐν μέσῳ κειμένων, τοῦ τε Λιβάνου καὶ τοῦ Ἀντιλιβάνου καὶ τοῦ Κασίου: οἱ μὲν δὴ τοιαῦτα λέγουσι. 723 Strabo indicates the author and context of the citation (᾿Ερατοσθένης... μνησθεὶς... “Eratosthenes, when he mentions...”) then proceeds to the citation in reported speech governed by φησὶ. He then signals his own 721 Brodersen, ‘Artemidorus [3]’, BNP; Banchich, ‘Artemidoros of Ephesos (438)’, BNJ.; Dueck (2000) 59–60. 722 Strabo 16.2.33: “τοιαύτη μὲν ἡ Φοινίκη. φησὶ δ᾽ Ἀρτεμίδωρος εἰς τὸ Πηλούσιον ἐκ μὲν Ὀρθωσίας εἶναι σταδίους τρισχιλίους ἑξακοσίους πεντήκοντα κατακολπίζοντι: ἐκ δὲ Μελαινῶν ἢ Μελανιῶν τῆς Κιλικίας τῶν πρὸς Κελένδεριν ἐπὶ μὲν τὰ μεθόρια τῆς Κιλικίας καὶ Συρίας χιλίους καὶ ἐνακοσίους: ἐντεῦθεν δ᾽ ἐπὶ τὸν Ὀρόντην πεντακοσίους εἴκοσιν: εἶτ᾽ ἐπὶ Ὀρθωσίαν χιλίους ἑκατὸν τριάκοντα.” (Such is Phoenicia. Artemidorus says that the distance from Pelusium to Orthosia is 3650 stadia, including the sinuosities of the gulfs; and from Melaenae, or Melaniae, in Cilicia near Celenderis, to the common boundaries of Cilicia and Syria, 1900 stadia; and thence to the Orontes river 520 stadia; and then to Orthosia 1130 stadia.) Modified from the Loeb trans. 723 Strabo 16.1.12: “Eratosthenes, when he mentions the lakes near Arabia, says that when the water is deprived of exits it opens up underground passages and through these flows underground as far as the country of Coelê- Syria, and that it is pressed up into the region of Rhinocolura and Mount Casius and forms the lakes and the pits there; but I do not know whether or not his statement is plausible; for the offshoots of the Euphrates which form the lakes near Arabia and the marshes are near the Persian Sea, but the isthmus which separates them is neither large nor rocky , so that it was more likely that the water forced its way into the sea in this region, whether underground or on the surface, than that it traversed a distance of more than six thousand stadia, through a country so waterless and dry , and that too when mountains intervene, I mean Mt. Libanus, Mt. Antilibanus, and Mt. Casius. Such, then, are the accounts of Aristobulus and Eratosthenes.” Loeb trans. 207/448 opinion (οὐκ οἶδα δ᾽ εἰ πιθανῶς εἴρηκεν, “I don’t know if what he says is likely”), provides his counter- arguments, and explicitly closes the discussion (οἱ μὲν δὴ τοιαῦτα λέγουσι, “These people say such things”). Here Strabo displays three common features of his citation style: use of reported speech with φησὶ, 724 occasional interjection of his own opinion with first person verbs, 725 and passing reference to anonymous authors. 726 As can be seen, Strabo cites and responds to his sources on a regular basis. At some times he cites authors for the information their work contains (such as his lengthy paraphrase of Aristobulus and his use of Nearchus, Orthagorus and Artemidorus) and at others cites them to contradict their assertions (such as his report of Polyclitus’ opinion on the Euphrates). The vast majority of sources he names are Hellenistic writers, usually either noted geographers in their own right or people who had travelled through the areas under discussion and written works drawing on that first-hand experience of the spaces. In the most complex source engagement in his description of Mesopotamia, he cites the opinions of several authors on naptha: Eratosthenes is cited directly , a collective group of unnamed authors are paraphrased, then Poseidonius is invoked and paraphrased to conclude the section. 727 Nevertheless, Strabo cites his sources far less than he 724 Citations with φησι and indirect speech: Strabo 16.1.13 (Polyclitus); 16.1.15 and 16.2.4 (Posidonius); 16.2.33 (Artemidorus); Strabo 16.3.7 (Nearchus). At 16.1.11, Strabo cites Aristobulus at the beginning of the passage (“φησὶ δ᾽ Ἀριστόβουλος τὸν Ἀλέξανδρον ... ἐπισκοπεῖν” (Aristobulus says that Alexander inspected...)) then because of the length of this passage, he later reminds the reader that he is reporting Aristobulus’ work with another φησίν (“he says”). At 16.3.5, Strabo cites Nearchus and Orthagoras jointly with εἰρέω then reiterates the citation with φησιν: εἰρήκασι καὶ Νέαρχος καὶ Ὀρθαγόρας νῆσον... κεῖσθαι... (“Nearchus and Orthagoras have said that an island... lies...”); see above, n.690. Strabo also uses εἰρέω to quote Eratosthenes directly at 16.1.15, see below , n.695. 725 At 16.1.13 (= BNJ 128 F5) Strabo concludes his citation with his own opinion on the truth of the part of Polyclitus’ statements: τὸ μὲν οὖν ὕστατον ῥηθὲν φανερῶς ἄτοπον... (“This latter comment is clearly wrong...”). The full fragment is quoted above, n.694. 726 Strabo 16.1.15. 727 First, Eratosthenes is cited using εἴρηκεν and direct speech with ὅτι: Strabo 16.1.15: “περὶ ἧς Ἐρατοσθένης μὲν οὕτως εἴρηκεν ὅτι ‘ἡ μὲν ὑγρὰ ἣν καλοῦσι νάφθαν, γίνεται ἐν τῇ Σουσίδι, ἡ δὲ ξηρὰ δυναμένη πήττεσθαι ἐν τῇ Βαβυλωνίᾳ: ταύτης δ᾽ ἐστὶν ἡ πηγὴ τοῦ Εὐφράτου πλησίον: πλημμύροντος δὲ τούτου κατὰ τὰς τῶν χιόνων τήξεις καὶ αὐτὴ πληροῦται καὶ ὑπέρχυσιν εἰς τὸν ποταμὸν λαμβάνει: ἐνταῦθα δὲ συνίστανται βῶλοι μεγάλαι πρὸς τὰς οἰκοδομὰς ἐπιτήδειαι τὰς διὰ τῆς ὀπτῆς πλίνθου.” (Concerning which Eratosthenes states that the liquid kind, which is called naphtha, is found in Susis, but the dry kind, which can be solidified, in Babylonia; and that there is a fountain of this latter asphalt near the Euphrates River; and that when this river is in flood at the time of the melting of the snows, the fountain of asphalt is also filled and 208/448 must have used them; his normal practice is not to cite. Only one of the citations in book 16 is directly quoted, the rest are reworded at least enough to be presented in indirect speech, if not completely reworded. Strabo’s reworking of material can be seen by comparing a story that he recounts in book 16 to the same story in Athenaios, who cites Posidonius as the source. First Athenaios’ version: οἶδα δὲ καὶ Ποσειδώνιον τὸν ἀπὸ τῆς Στοᾶς εἰπόντα [καὶ] περὶ πλήθους ἰχθύων τάδε· “ὅτε Τρύφων ὁ Ἀπαμεὺς ὁ τὴν τῶν Σύρων βασιλείαν ἁρπάσας ἐπολεμεῖτο ὑπὸ Σαρπηδόνος τοῦ Δημητρίου στρατηγοῦ περὶ Πτολεμαίδα πόλιν καὶ ὡς ὁ Σαρπηδὼν λειφθεὶς 1 ἀνεχώρησεν εἰς τὴν μεσόγαιαν μετὰ τῶν ἰδίων στρατιωτῶν, οἱ δὲ τοῦ Τρύφωνος ὥδευον κατὰ τὸ πλησίαλον νικήσαντες τῆι μάχηι, ἐξαίφνης πελάγιον κῦμα ἐξαρθὲν μετέωρον εἰς ὕψος ἐξαίσιον ἐπῆλθεν τῆι γῆι καὶ πάντας αὐτοὺς ἐπέκλυσεν διέφθειρέν τε ὑποβρυχίους, ἰχθύων τε πολὺν σωρὸν ἀναχωροῦν τὸ κῦμα μετὰ τῶν νεκρῶν κατέλιπε. καὶ οἱ περὶ τὸν Σαρπηδόνα ἀκούσαντες τὴν συμφορὰν ἐπελθόντες τοῖς μὲν τῶν πολεμίων σώμασιν ἐφήσθησαν, ἰχθύων δὲ ἀφθονίαν ἀπηνέγκαντο καὶ ἔθυσαν Ποσειδῶνι τροπαίωι πρὸς τοῖς προαστείοις τῆς πόλεως.” ... 728 Athenaios claims to be quoting Posidonius directly with ὅτε and direct speech. Strabo reports the same story indirectly and without citation: overflows into the river; and that there large clods of asphalt are formed which are suitable for buildings constructed of baked bricks.) Loeb trans. Second, The opinions of unnamed “others” are expressed using indirect speech. Strabo 16.1.15: “ἄλλοι δὲ καὶ τὴν ὑγρὰν ἐν τῇ Βαβυλωνίᾳ γίνεσθαί φασι. περὶ μὲν οὖν τῆς ξηρᾶς εἴρηται πόσον τὸ χρήσιμον τὸ ἐκ τῶν οἰκοδομιῶν μάλιστα: φασὶ δὲ καὶ πλοῖα πλέκεσθαι, ἐμπλασθέντα δ᾽ ἀσφάλτῳ πυκνοῦσθαι... πείρας δὲ χάριν φασὶν Ἀλέξανδρον ἐν λουτρῷ προσχέαι παιδὶ τοῦ νάφθα καὶ προσαγαγεῖν λύχνον...” (Other writers say that the liquid kind also is found in Babylonia. Now writers state in particular the great usefulness of the dry kind in the construction of buildings, but they say also that boats are woven with reeds and, when plastered with asphalt, are impervious to water... It is said that Alexander, for an experiment, poured some naphtha on a boy in a bath, and brought a lamp near him...) Loeb trans. Finally , Posidonius concludes the passage in indirect speech. Strabo 16.1.15: “Ποσειδώνιος δέ φησι τοῦ ἐν τῇ Βαβυλωνίᾳ νάφθα τὰς πηγὰς τὰς μὲν εἶναι λευκοῦ τὰς δὲ μέλανος...), τὰς δὲ τοῦ μέλανος ἀσφάλτου ὑγρᾶς, ᾧ ἀντ᾽ ἐλαίου τοὺς λύχνους κάουσι.” (Poseidonius says of the springs of naphtha in Babylonia, that some send forth white naphtha and others black...) Loeb trans. 728 BNJ 87 F 29 = Ed.-Kidd F 226 = Athenaios, Deipnosophists 8.7: “And I know that Poseidonios the Stoic, too, said the following about a huge number of fish: ‘Tryphon of Apameia, the man who seized the throne of Syria, was attacked by Sarpedon, the general of Demetrios, near the city of Ptolemais, and Sarpedon, defeated, withdrew inland with his own troops, whilst those of Tryphon, victorious in the battle, took the coastal road. Suddenly a wave from the sea, raised up to an extraordinary height, invaded the land, and engulfed them all. It drowned them and the wave withdrew , it left behind a huge heap of fish together with the bodies. Sarpedon’s men heard about the disaster and when they arrived on the scene they were delighted at the bodies of their enemies, carried off an abundance of fish, and sacrificed to Poseidon Tropaios at the suburbs of the city.’ ...” Dowden trans. 209/448 Ἱστορεῖται δὲ παράδοξον πάθος τῶν πάνυ σπανίων κατὰ τὸν αἰγιαλὸν τοῦτον τὸν μεταξὺ τῆς τε Τύρου καὶ τῆς Πτολεμαΐδος. καθ’ ὃν γὰρ καιρὸν οἱ Πτολεμαεῖς μάχην συνάψαντες πρὸς Σαρπηδόνα τὸν στρατηγὸν ἐλείφθησαν ἐν τῷ τόπῳ τούτῳ τροπῆς γενομένης λαμπρᾶς, ἐπέκλυσεν ἐκ τοῦ πελάγους κῦμα τοὺς φεύγοντας ὅμοιον πλημμυρίδι, καὶ τοὺς μὲν εἰς τὸ πέλαγος ἀπήρπασε καὶ διέφθειρεν, οἱ δ’ ἐν τοῖς κοίλοις τόποις ἔμειναν νεκροί• διαδεξαμένη δὲ ἡ ἄμπωτις πάλιν ἀνεκάλυψε καὶ ἔδειξε τὰ σώματα τῶν κειμένων ἀναμὶξ ἐν νεκροῖς ἰχθύσι ... 729 These stories both reflect the same incident from the mid-second century BCE. 730 Strabo’s version of the event refers to a source obliquely (Ἱστορεῖται δὲ παράδοξον, “a story is told of an amazing event”). Although Strabo does not quote a source for the passage, the similarity with the directly quoted fragment found in Athenaios suggests that a common source is likely in Poseidonius. The stories include many of the same narrative features: a geographical reference to the location of the event (1), 731 the main protagonists (2), a huge wave of the sea presented in similar language (3), and a pile of bodies mixed with fishes (4). Feature Poseidonius Strabo 1 περὶ Πτολεμαίδα πόλιν τὸν μεταξὺ τῆς τε Τύρου καὶ τῆς Πτολεμαΐδος 2 Τρύφων ὁ Ἀπαμεὺς ὁ τὴν τῶν Σύρων βασιλείαν ἁρπάσας ἐπολεμεῖτο ὑπὸ Σαρπηδόνος τοῦ Δημητρίου οἱ Πτολεμαεῖς μάχην συνάψαντες πρὸς Σαρπηδόνα τὸν στρατηγὸν 3 ἐξαίφνης πελάγιον κῦμα ἐξαρθὲν μετέωρον εἰς ὕψος ἐξαίσιον ἐπῆλθεν τῆι γῆι καὶ πάντας αὐτοὺς ἐπέκλυσεν διέφθειρέν τε ὑποβρυχίους ἐπέκλυσεν ἐκ τοῦ πελάγους κῦμα τοὺς φεύγοντας ὅμοιον πλημμυρίδι 4 διέφθειρέν τε ὑποβρυχίους, ἰχθύων τε πολὺν σωρὸν ἀναχωροῦν τὸ κῦμα μετὰ τῶν νεκρῶν κατέλιπε διαδεξαμένη δὲ ἡ ἄμπωτις πάλιν ἀνεκάλυψε καὶ ἔδειξε τὰ σώματα τῶν κειμένων ἀναμὶξ ἐν νεκροῖς ἰχθύσι 729 Strabo 16.2.26: “An amazing event is told, which has very few parallels, and took place on this sea shore between Tyre and Ptolemaïs. On an occasion when the Ptolemaians had joined battle with the general Sarpedon and remained in this place following a brilliant rout (of Sarpedon), a wave from the sea like a flood-tide overwhelmed them as they fled (from it). Some were carried off into the sea and killed; others remained in depressions, dead. As the sea receded again, it uncovered and revealed the bodies of the fallen with fish mixed in with the corpses ...” Dowden trans. 730 Either 139 BCE (Dowden, BNJ 87 F 29 Commentary) or 144/3 BCE (Ed.-Kidd F 226 Comment, Kidd (1989) 808). 731 On Ptolemais, see Cohen (2006) 213–21. 210/448 The latter two elements preserve respectively the vocabulary and an evocative metaphor of Poseidonius. While Poseidonius (as preserved in Athenaios) painted a vivid picture of the event, including the powerful image of bodies mixed with fishes, his report is somewhat matter-of-fact when compared to the more stylised and literary description of Strabo. Moreover, Strabo omits certain precise aspects of the story which he judged superfluous to his geographical description of the Syria coast. Strabo elides the the name of the victorious general, Tryphon of Apameia, as well as the more precise notes on the movement of the two armies that Poseidonius provided, including the return of Sarpedon’s army , the collection of the fish and the sacrifice to Poseidon Tropaios. From this case we can see that even when Strabo does not report a source, he is selecting and reworking material from one or more of his many sources and creating his own narrative from that material. The inclusion and exclusion of any given geographical feature, description or measurement was his own choice. He is dependent on the information that his sources provide, but he is not bound by their selective choices. The geography he constructs is therefore to some extent his own. Why then does he periodically choose to mention the names of his sources at all? There is no clear pattern to the occasions on which he reports them. He does not exclusively use citations to present a position with which he intends to disagree. While he sometimes disagrees, refutes or casts doubt on the opinions of those he cites, he equally often gives no indication that the cited opinion should be disregarded. Nor is there any indication that he cites sources in order to preserve any particular syntactical features. The answer lies at the start of his Geography . At 1.2.1, Strabo notes that it is noble (καλόν) to engage with worthy predecessors like Eratosthenes, Hipparchus, Posidonius and Polybius and such others (ἄλλους τοιούτους). 732 His use of these authors is thus a claim to 732 Strabo 1.2.1: ὥστε ἔχοιμεν ἄν τι λέγειν πλέον τῶν πρὸ ἡμῶν. ὁρᾶν δ’ ἔσται τοῦτο μάλιστα ἐν τοῖς λόγοις τοῖς πρὸς τοὺς πρὸ ἡμῶν, ἧττον μὲν τοὺς πάλαι, μᾶλλον δὲ τοὺς μετ’ Ἐρατοσθένη καὶ αὐτὸν ἐκεῖνον· οὓς εἰκὸς ὅσῳπερ πολυμαθέστεροι τῶν πολλῶν γεγόνασι, τοσούτῳ δυσελεγκτοτέρους εἶναι τοῖς ὕστερον, ἄν τι πλημμελῶς λέγωσιν. εἰ δ’ ἀναγκασθησόμεθά που τοῖς αὐτοῖς ἀντιλέγειν, οἷς μάλιστα ἐπακολουθοῦμεν κατ’ ἄλλα, δεῖ συγγνώμην ἔχειν· οὐ γὰρ πρόκειται πρὸς ἅπαντας ἀντιλέγειν, ἀλλὰ τοὺς μὲν πολλοὺς ἐᾶν, οἷς μηδὲ ἀκολουθεῖν ἄξιον, ἐκείνους δὲ διαιτᾶν, οὓς ἐν τοῖς πλείστοις κατωρθωκότας ἴσμεν, ἐπεὶ οὐδὲ πρὸς 211/448 authority . By naming such notable scholars he validates and authorises his own scholarship and the geography he constructs within it. There is no evidence that Strabo travelled in the region of Mesopotamia, so he relied on the reports of others in composing his Geography. As far as we can tell, Strabo used the best sources available to him. Of the seven sources Strabo names for book 16, four were witnesses to Alexander’s campaigns and the other three were geographers who probably had access to first-hand reports. Eratosthenes, Strabo’s most important source, certainly had access to the literary resources of Alexandria. 733 Moreover, Strabo was heir to the rigorous geographical debates among Hellenistic scientific geographers like Eratosthenes, Hipparchus and Poseidonius. 734 This is not to say that Strabo’s work was based only on Hellenistic Greek sources and relied solely on outdated material. Strabo displays his familiarity with contemporary events at several points in his work. His historical research included Roman activities in Syria and Mesopotamia. 735 As a native of Pontus, Mithridates was of particular interest to him and he included events from the Roman campaigns against Mithridates and Tigranes in the 80s, 70s and 60s BCE. 736 Strabo connected several sites in Syria to the Parthian invasions of the 40s BCE, included several Tiberian political perspectives and noted the first ἅπαντας φιλοσοφεῖν ἄξιον, πρὸς Ἐρατοσθένη δὲ καὶ Ἵππαρχον καὶ Ποσειδώνιον καὶ Πολύβιον καὶ ἄλλους τοιούτους καλόν.” (I may therefore have something more to say than my predecessors. This will become particularly apparent in what I shall have to say in criticism of my predecessors, but my criticism has less to do with the earliest geographers than with the successors of Eratosthenes and Eratosthenes himself. For it stands to reason that because Eratosthenes and his successors have had wider knowledge than most geographers, it will be correspondingly more difficult for a later geographer to expose their errors if they say anything amiss. And if I shall, on occasion, be compelled to contradict the very men whom in all other respects I follow most closely , I beg to be pardoned; for it is not my purpose to contradict every individual geographer, but rather to leave the most of them out of consideration—men whose arguments it is unseemly even to follow—and to pass upon the opinion of those men whom we recognise to have been correct in most cases. Indeed to engage in philosophical discussion with everybody is unseemly , but it is honourable to do so with Eratosthenes, Hipparchus, Poseidonius, Polybius, and others of their type.) Loeb trans. 733 Blomqvist (1992). 734 Hübner, Hipparchus [6], BNP ,; Viedebantt (1915). 735 The rest of Strabo’s Geography is sprinkled with commentary on recent Roman events, in particular Aelius Gallus’ expedition into Arabia (16.4.22-24), the influence of Roman control on Egyptian commerce (17.1.11- 12), Strabo’s judgements of the Roman empire (6.4.2; 17.3.24-25). 736 On Strabo’s Pontic background: Dueck (2000) 3–5. 212/448 annexation of Commagene in 18 CE. 737 Strabo thus depended on Hellenistic Greek geographical works as the basis of his Geography , but selectively created his own narrative by mixing and rearranging that material with more recent knowledge obtained from Roman sources. These factors all contribute to Strabo’s creation of a geography of the world centred on Rome. 738 This focus will be further explored in Chapter 4. Pliny’s Sources As a Roman veteran, a scholar and an administrator from an Italian family with close links to the imperial house, we might expect Pliny to display a different approach to knowledge production and source use from Strabo, and in fact he does to an extent. However, the question of Pliny’s geographical sources is much more complicated than Strabo’s, because of both the way he reports his sources and his overall geographical project. This section begins with Pliny’s overt statements about his sources and his methodology , then shows how the reality of Pliny’s production of geographical knowledge differed from his explicit statements. Like Strabo, Pliny relied on both autopsy and literary reports in his descriptions of the world and its contents. His discussion of the nature and variety of winds in book 2 includes a contrast between the two modes of knowledge acquisition. Viginti amplius auctores Graeci veteres prodidere de his observationes. Quo magis miror orbe discordi et in regna, hoc est in membra, diviso tot viris curae fuisse tam ardua inventu, inter bella praesertim et infida hospitia, piratis etiam, omnium mortalium hostibus, transituros fama terrentibus, ut hodie quaedam in suo quisque tractu ex eorum commentariis, qui numquam eo accessere, verius noscat quam indigenarum scientia, nunc vero pace tam festa, tam gaudente proventu rerum artiumque principe, omnino nihil addisci nova inquisitione, immo ne veterum quidem inventa perdisci. [118] Non erant maiora praemia, in multos dispersa fortunae magnitudine, et ista plures sine praemio alio quam posteros iuvandi eruerunt. Namque mores hominum senuere, non fructus, et inmensa multitudo aperto, quodcumque est, mari hospitalique litorum omnium adpulsu navigat, sed lucri, non scientiae, gratia. Nec reputat caeca mens et tantum avaritiae intenta id ipsum scientia posse tutius fieri. Quapropter scrupulosius, quam instituto fortassis conveniat operi, tractabo ventos, tot 737 Pothecary (2002). 738 Dueck (2000) 101–106. 213/448 milia navigantium cernens. 739 The temporal context of Pliny’s text is confused, but he contrasts between a time of wars, treachery and piracy (hodie) and a time of peace and potential scientific discovery (nunc). In the former time, widespread discord means that better knowledge is available from the writing of others (ex eorum commentariis) than from the knowledge of people native to the space in question (indigenarum scientia). In the latter time, peace allows an immense degree of maritime traffic, but those undertaking it are more concerned with commerce than the production and transmission of knowledge. In fact, this is not so much a contrast between the use of first- and second-hand knowledge, as it is a contrast between the means of transmission of second-hand knowledge to the geographical writer. Pliny discloses no desire to travel to see things himself; rather he wishes those people already travelling would be more concerned with knowledge and bring that knowledge back to him. Second-hand knowledge is adequate for the geographic writer, and perhaps even preferred to personal travel. 740 This fits well with the literary focus of Pliny’s work. Pliny was a well-read scholar who listed a formidable range of sources in his first book. That book functions as a table of contents for the rest of the work; he lists the subject material covered by each book and 739 Pliny NH 2.117-118: “More than twenty old Greek writers have published their observations about these [winds]. And this is the more remarkable, seeing that there is so much discord in the world, and that it is divided into different kingdoms, that is into separate members, that there should have been so many who were concerned with such difficult investigations, especially at a time of wars and treachery , and when pirates, the enemies of everyone, terrify with their infamy everyone who moves about, so that today a person may acquire truer knowledge about a country from the writings of those who have never been there, than from the inhabitants themselves. But now in a time of such joyous peace which, under a princeps who so greatly encourages the growth of the arts, no new inquiries have been made at all, nor are the discoveries of the ancients mastered. There were not greater rewards, with the great prosperity being spread among many people, but that there were more people who studied these things seeking no reward except benefiting posterity. For the morals of men have grown weak, not the rewards, and since all the seas, however many there are, lie open and every shore one approaches is hospitable, a great number of people undertake voyages; for the sake of profit, not knowledge. Nor does the blinded mind intent on greed reckon that this very thing might be able to be done more safely by knowledge. In which case, since there are so many people sailing, I will discuss the winds in more careful detail than might seem appropriate to the work I am constructing.” 740 Pliny does occasionally refer to his own first-hand experience of certain natural phenomena. Some examples are discussed by Lloyd (1983) 136–38; Murphy (2004) 8–11. 214/448 the sources from which his information derives. 741 That the entire first book is devoted to a lengthy accounting of his literary sources is in itself a powerful argument for Pliny’s prioritisation of secondary knowledge. He lists his sources for each of the books in his encyclopaedia separately; divided into Roman and foreign. For book 5, which covers Africa and Asia, he provides the following: Ex Auctoribus: Agrippa. Suetonio Paulino. M. Varrone. Varrone Atacino. Cornelio Nepote. Hygino. L. V etere. Mela. Domitio Corbulone. Licinio Muciano. Claudio Caesare. Arruntio. Livio filio. Seboso. actis triumphorum. Externis: Iuba rege. Hecataeo. Hellanico. Damaste. Dicaearcho. Baetone. Timosthene. Philonide. Xenagora. Astynomo. Staphylo. Dionysio. Aristotele. Aristocrito. Ephoro. Eratosthene. Hipparcho. Panaetio. Serapione Antiocheno. Callimacho. Agathocle. Polybio. Timaeo mathematico. Herodoto. Myrsilo. Alexandro polyhistore. Metrodoro. Posidonio qui περιπλουν aut περιηγησιν. Sotade. Pyrrandro. Aristarcho Sicyonio. Eudoxo. Antigene. Callicrate. Xenophonte Lampsaceno. Diodoro Syracusano. Hannone. Himilcone. Nymphodoro. Calliphane. Artemidoro. Megasthene. Isidoro. Cleobulo. Aristocreonte. 742 Neither section is arranged chronologically and both contain a wide variety of geographical, historical and scientific writers. Pliny includes many of Strabo’s sources for book 16 in this list (Eratosthenes, Poseidonius (Posidonio qui περιπλουν aut περιηγησιν) and Artemidorus) as well as other sources which Strabo used elsewhere (Hipparchus, Polybius and Megathenes). However, in his geographical books, Pliny seldom refers to his sources by name. As we shall see, where he does cite prior authorities, he prefers Romans or writers closely associated to the Romans, such as the Roman client king Juba (Iuba rege). In fact, rather than Roman and foreign, it would be more accurate to say that Pliny divides the authors on which he relies into sources (ex auctoribus), and foreign sources (externis), where Roman authority is preferred and that ex auctoribus externis is relegated to subordinate status. This organisational 741 The omission of Strabo is usually taken as evidence that his Geography was not widely known in the first century CE; however, another absence must cast doubt on that supposition. W oolf (2011) 10 notes the surprising omission of Sallust from Pliny’s sources on Africa; there were certainly aspects of the Bellum Iugurthinum relevant to Pliny’s presentation of the African provinces, even if Sallust’s geographical presentation in that work was at times oddly incorrect (for example, BI 23.1 with Paul (1984) 83 and BI 92.3 with Paul (1984) 230; see also the brief note at Syme (1964) 152. 742 Pliny , NH 1, Libro V . 215/448 principle is indicative of the work as a whole in which the world as conquered and organised by Roman authority is displayed before the eyes of the Imperial family , specifically the future princeps Titus, to whom the work is dedicated. 743 Pliny’s index is an impressive array of scholarship, in keeping with the younger Pliny’s claims about his uncle’s working habits. 744 However, perhaps more important than their geographical writings, these names served as a powerful authorising claim for Pliny’s work. Note in particular a number of sources at the very heart of the imperial structure: Marcus Agrippa, right hand man of Augustus; the emperor Claudius; Domitius Corbulo, Nero’s great general; Licinius Mucianus, Vespasian’s amicus who seized and governed Rome on his behalf in the civil war of 69 CE. Pliny also names the triumphal archives (actis triumphorum) as a geographical source. The sources for his sixth book are similarly impressive. 745 Pliny himself was firmly embedded in Roman imperial structures. 746 His equestrian career included military service on the Rhine frontier and probably multiple provincial procuratorships. 747 This experience gave him ample first hand knowledge in the means and reality of Roman power at both the centre and periphery of the empire. 748 He spent his otium in writing works in support of the same imperial structures, as seen in his own 743 Pliny , Praefatio; Murphy (2004) 164. Murphy also argues that the index has an objectifying function: “Not only does the index announce the encyclopedia's universal scope and the laborious assembly of its sources, it is most impressive for showing how totally knowledge can be made into an object. It is a concrete demonstration of how referential shape can be imposed on the whole of nature, and a successful demarcation of the world into class and subclass: here everything is made indexible. The index is a textual analogy to the Natural History's use of the Roman triumphal procession as heuristic and analytical metaphor: both triumph and index make a thing known while at the same time signifying both the total availability and the total objectification of what is known.” (p.214). 744 Pliny , Ep. 3.5. But note Pliny the Younger’s construction of Pliny the Elder, see W oolf (2011) n.14. 745 Pliny , NH 1, Libro VI: Ex Auctoribus: M. Agrippa. M. Varrone. Varrone Atacino. Cornelio Nepote. Hygino. L. V etere. Mela Pomponio. Domitio Corbulone. Licinio Muciano. Claudio Caesare. Arruntio. Seboso. Fabricio Tusco. T . Livio filio. Seneca. Nigidio. Externis: Iuba rege. Hecataeo. Hellanico. Damaste. Eudoxo. Dicaearcho. Baetone. Timosthene. Patrocle. Deomdamante. Clitarcho. Eratosthene. Alexandro Magno. Ephoro. Hipparcho. Panaetio. Callimacho. Artemidoro. Agathocle. Polybio. Timaeo Siculo. Alexandro polyhistore. Isidoro. Amometo. Metrodoro. Posidonio. Onesicrito. Nearcho. Megasthene. Diogneto. Aristocreonte. Bione. Dalione. Simonide minore. Basile. Xenophonte Lampsaceno. 746 Murphy (2004) 2. 747 Pliny’s military service in Germany: Syme (1969) 204–8. Syme discusses the evidence for the four procuratorships which have been proposed for Pliny (Narbonensis, Gallica Belgica, Africa and Tarraconensis), Syme (1969) 208–18; 224–27. Evidence for the first two is very weak, and that for Africa not much better; Tarraconensis seems secure. 748 Murphy (2004) 5. 216/448 writings and the testimony of his nephew . 749 This literary production adhered to a Roman logic of aristocratic literary exchange that saw the exchange of knowledge as a morally good demonstration of greatness and social alliance. 750 In Pliny’s view , knowledge should be a shared possession of Roman citizens which contributes to their mastery of the environment. 751 His Natural History was as much a part of an imperialistic literary output as any of his works. 752 How , then, did Pliny use these sources? In his preface, Pliny discusses his engagement with his predecessors and how they used their source material: argumentum huius stomachi mei habebis quod in his voluminibus auctorum nomina praetexui. est enim benignum, ut arbitror, et plenum ingenui pudoris fateri per quos profeceris, non ut plerique ex iis, quos attigi, fecerunt. [22] scito enim conferentem auctores me deprehendisse a iuratissimis ex proximis veteres transcriptos ad verbum... [23] Obnoxii profecto animi et infelicis ingenii est deprehendi in furto malle quam mutuum reddere, cum praesertim sors fiat ex usura. 753 Despite these objections in his preface, Pliny himself seldom refers to his sources in the geographical sections of his work. In the section on Syria (NH 5.66-90), he refers to his sources on two occasions: Herodotus and other unnamed writers (“quidam”) on the dimensions and location of the Sirbonian lake (Sirbonis lacus; now Sabkhat al Bardawil on the north coast of the Sinai peninsula); 754 and Domitius Corbulo and Licinius 749 Pliny the Younger lists his uncles works in Ep. 3.5. They included a biography of a consular partron, and works on history , military tactics, and the education of orators. Further biographical details can be found in Ep. 6.16, 6.20. 750 Murphy (2004) 49–73. For examples of knowledge as a gift, see Cic. Brutus (13-19); Pliny Ep. 9.33 (who gives a story about a dolphin which he took from his uncle’s work (NH 9.26-8); Pliny Ep. 6.16 (the story of his uncles death as a story given to Tacitus). 751 Pliny NH 25.1-2, 24.5; French (1994) 206–8. 752 Murphy (2004) 11: “The Natural History was never a pure vessel for transmitting knowledge, created in a vacuum for its own sake. It was an instance of knowledge given institutional form, for a purpose, created under particular social and political circumstances, and it still bears the marks of that creation.” 753 Pliny NH. Pr.21-23: “You have evidence of my taste in that I have placed the names of my sources at the start of this volume. For in my opinion it is a good practice and full of noble modesty to acknowledge those through whom one makes progress, not to act as most of those I have come across do. [22] For I must let you know that I have compared authors and discovered from the most solemn recent authors have copied older works word for word... [23] Certainly it takes a servile spirit and and ill-omened nature to prefer to be discovered in theft than to return a loan, especially when capital has been acquired from the interest.” 754 Pliny NH. 5.68. 217/448 Mucianus on the source of the Euphrates. 755 In the section on Mesopotamia (NH 6.117-134), Pliny mentions Nearchus, Onesicritus, Juba and other unnamed authorities, including Persian writers, on the length of the Euphrates. 756 Later, he describes Claudius Caesar’s account of the behaviour of a tributary of the Tigris. 757 Finally , amidst his description of Babylonia, Pliny gives the dimensions and relative location of Media, Parthia, Persis and Mesopotamia according to Marcus Agrippa. 758 Pliny’s model of aristocratic literary exchange placed citations of fellow Roman aristocrats at a higher value than others, and certainly above first- hand knowledge. Moreover, rather than specify the individual sources of each piece of his compiled geographical knowledge, Pliny credits them collectively (with occasional exceptions, often in cases of disagreement among his authorities), probably concealing just how much of his knowledge was derived ex auctoribus externis. Of these cited sources, all are mentioned in the list in NH 1, except for Nearchus and Onesicritus. They are cited for the length of the Euphrates; perhaps Pliny found this citation in another of his sources. 759 Fragments providing distance information frequently occur like this in ancient geographical writing, as Pliny himself shows here and as can also been seen on many occasions in Strabo. 760 As with many of the Roman geographical writers, earlier scholars were sceptical of his methods; Syme noted Pliny’s wide 755 Pliny NH. 5.83. As well as an important figure in contemporary politics, Licinius Mucianus (cos. III 74) was an major source on eastern marvels, Murphy (2004) 203–4. 756 Pliny , NH. 6.124: “Euphrate navigari Babylonem e Persico mari CCCCXII [m.]p. tradunt Nearchus et Onesicritus; qui vero postea scripsere, ad Seleuciam CCCCXL [m.p.]; Iuba a Babylone Characen CLXXV D [m.p.]. fluere aliqui ultra Babylona continuo alveo, prius quam distrahitur ad rigua, LXXXVII [m.p.], universo autem cursu XII [m.]p. inconstantiam mensurae diversitas auctorum facit, cum Persae quoque schoenos et parasangas alii alia mensura determinent.” 757 Pliny , NH. 6.128. 758 Pliny , NH. 6.136-7. 759 Pliny , NH 6.96 = BNJ 133 F 13 suggests that Pliny did not consult Nearchus directly. Pliny says that Nearchus did not include staging posts and distances, but he did. (Whitby , ‘Nearchus (133)’, BNJ). Two other sources cited by Strabo also appear elsewhere in Pliny’s Natural History. Poseidonios is cited by Pliny as an authority for books 2, 4, 5 (no explicit references), 6? (cited at 6.57), 11. Polyclitus of Larisa (4th C BCE) appears in Pliny’s sources for books 12, 13, 31 (in the citation for books 12 and 13, he appears as Polykritos, which Jacoby considered a mistake, perhaps by Pliny himself, FGrH 128 T 2a ; see also Plutarch, F 8; Pearson (1960) 70–77. 760 In addition to those already mentioned, see, for example, Strabo 16.3.5. 218/448 reading, but criticised him for hasty and insufficiently selective use of his sources. 761 However, recent works examining the project of the Natural History more closely have been more charitable. In his volume on Roman ethnographic thought, W oolf finds Pliny to be a deliberate and selective compiler and arranger of material. 762 This can be seen in his arrangement of disparate geographic sources to form a coherent narrative of the Roman world. 763 Nevertheless, despite the Roman focus of Pliny’s work, he was dependent on those foreign authorities for much of his information about the world, in particular the eastern regions. These had been comparatively well-documented by Hellenistic writers, but relatively recently entered the sphere of Roman interest and scholarship. Pliny refers to various Greek authors throughout his work, but he asserts the novelty of his own encyclopaedic project in his preface, noting there that no Greek had written an encyclopaedia. 764 He goes on to deride the flippant names and trivial contents of Greek works; the first of many jabs at Greek scientific thought. 765 Nevertheless, his work attests to his reliance on Greek sources in every field, as the chapters of a conference volume on Plinian science attest. 766 Aristotle, Theophrastus and Democritus were particularly important. 767 Pliny’s work was certainly influenced by Stoic philosophical writings, especially on matters of physics, although his conception of the universe as aeternum, inmensum, neque genitum neque interiturum umquam reveals that Stoicism was not his only source. 768 Pliny rarely refers to these Greek sources of 761 Syme (1969) 202. 762 W oolf (2011) 11. 763 W oolf (2011) 11. In all, Pliny’s use of sources resembles that of a contemporary work of popular history where the sources are hidden behind the text and seldom explicitly cited. 764 Pliny NH Pr.14. 765 Pliny NH Pr.24; French (1994) 218–19; on doctors, 223–25; and in a similar vein on Magi, 225–28.. W allace- Hadrill shows that Pliny’s attitude towards Greek and Roman modes of scientific thought are closely related to his presentation and moralising on the antithesis between luxury and Nature, (1990) 92–96. 766 French and Greenaway (1986). 767 French (1994) 220–23. 768 Pliny NH 2.1: “permanent, boundless, which was neither created, nor will ever be destroyed”. Lapidge (1989) 1411–12; Wallace-Hadrill (1990) 84; Beagon (1992) 26–54; French (1994) 198–201; Paparazzo (2011). Healy (1999) 71–78, 106–112. For the Stoic conception of a cyclical universe ending and reborn in destruction, see the sources assembled by Long and Sedley (1987) 1.274–79, 2.271–77. 219/448 knowledge in his geographical sections, but the material which comprises his descriptions of Mesopotamia reveals his dependence on those sources. 769 In a sense, Pliny treats his foreign geographical sources as assumed knowledge not requiring precise citation. The specific details of earlier geographic research had become a possession of Roman scholarship. In this, Pliny’s narrative geography is similar to Ptolemy’s scientific tradition. Ptolemy’s Sources Whereas the other extant Roman imperial geographic writers created works of narrative geography , Ptolemy participated in the scientific tradition associated with Alexandrian scholarship. His cultural background is unclear: he bore a Macedonian name of considerable prestige in an Egyptian context and worked in a multicultural metropolis at the height of the Roman empire. His major source was Marinus of Tyre. From his toponymic, Marinus’ origin was the region of Phoenicia, but Ptolemy is silent on further details. His name was probably Latin. 770 Ptolemy’s choice of Greek to transmit his scholarship probably indicates his adherence to the Hellenistic scientific tradition, as does his method. Ptolemy constructed his work through the accumulation of knowledge, explicitly correcting and supplementing the work of his predecessor Marinus, rather than substituting and selectively arranging material from a variety of sources to construct space. To the extent that Marinus did the same, we can speak of Ptolemy’s participation in a tradition of knowledge: a process which prioritised the accretion and sedimentation of geographical facts rather than a selective process of narrative construction. Before embarking on a detailed exploration of the accumulation of knowledge through sedimentation, I shall discuss Ptolemy’s explicit attitude towards the sources of that knowledge. Ptolemy is forthright about his major source, the late first century CE geographer Marinus of Tyre, 769 See below , pp.232ff. 770 Although Dilke (1985) 72. suggests it could have been Greek, Latin, or Aramaic. 220/448 although that scholar’s sources are obscure. 771 Δοκεῖ δὴ Μαρῖνος ὁ Τύριος ὕστατός τε τῶν καθ’ ἡμᾶς καὶ μετὰ πάσης σπουδῆς ἐπιβαλεῖν τῷ μέρει τούτῳ· φαίνεται γὰρ καὶ πλείοσιν ἱστορίαις περιπεπτωκὼς παρὰ τὰς ἔτι ἄνωθεν εἰς γνῶσιν ἐλθούσας, καὶ τὰς πάντων σχεδὸν τῶν πρὸ αὐτοῦ μετ’ ἐπιμελείας διειληφώς, ἐπανορθώσεώς τε τῆς δεούσης ἀξιώσας, ὅσα μὴ προσηκόντως ἐτύγχανε πεπιστευμένα καὶ ὑπ’ ἐκείνων καὶ ὑφ’ ἑαυτοῦ τὸ πρῶτον, ὡς ἐκ τῶν ἐκδόσεων αὐτοῦ τῆς τοῦ γεωγραφικοῦ πίνακος διορθώσεως πλειόνων οὐσῶν ἔνεστι σκοπεῖν. 772 Ptolemy notes that Marinus had many sources and implies a process of selection and aggregation among them. He does not discuss the nature of these sources, but his implicit statement that the geographic writer must place trust in them (πεπιστευμένα) recalls Strabo’s distinction between information gathered through autopsy and that gathered through trusting (πιστεύσαντες). 773 Ptolemy does not explicitly develop the contrast between these sources of geographic knowledge (although evidence that Marinus used both will be discussed below), although he does imply that this trust is fallible. Ptolemy explains that Marinus’ method was primarily corrective (ἐπανορθώσεώς) because sometimes a geographical writer would place unsuitable trust in their sources of information (ὅσα μὴ προσηκόντως ἐτύγχανε πεπιστευμένα καὶ ὑπ’ ἐκείνων καὶ ὑφ’ ἑαυτοῦ τὸ πρῶτον). Marinus corrected both the work of his predecessors (ὑπ’ ἐκείνων) and his own earlier work (ὑφ’ ἑαυτοῦ τὸ πρῶτον). Ptolemy’s knowledge that Marinus’ method included self-correction suggests that Ptolemy had access to several iterations of Marinus’ publications. Ptolemy notes that Marinus revised his “map” (πίναξ) many times (τῆς τοῦ γεωγραφικοῦ πίνακος διορθώσεως πλειόνων οὐσῶν). It is unclear what form this map took; the maps which accompany Ptolemy’s manuscripts were probably drawn from Ptolemy’s data in 771 Gärtner, ‘Marinus’, BNP; Berggren and Jones (2000) 23–25; Dilke (1985) 72–75, 77. 772 Ptol. Geog. 1.6.1: “Marinus of Tyre seems to be the latest [author] in our time to have undertaken this subject, and he has done it with absolute diligence. He has clearly laid his hands on numerous records of research besides those that had come to knowledge still earlier, and treated those of nearly all his predecessors with care, giving appropriate correction to everything that he found that either they or he himself, at first, had trusted without good reason, as can be seen from his publications of the geographical map [τοῦ γεωγραφικοῦ πίνακος], which are numerous.” Berggren (2000) trans. 773 Strabo 2.5.11, quoted and discussed above, p.197. 221/448 the Middle Ages. 774 Ptolemy states that Marinus did not have time to draw a map to accompany the final version of his work. 775 From Ptolemy’s subsequent statements that drawing such a map would have been the only way to correct certain errors which Ptolemy goes on to explain and correct, we may assume that Ptolemy did draw such a map, but if it was published with his work, the original has been lost. Ptolemy may be using πίναξ in the sense of a register or list (Lat. album) to refer to Marinus’ catalogue of geographical coordinates rather than in the sense of a completed map with those same sites located in a two-dimensional plane. If so, Marinus’ πίναξ was a guide to constructing a projected map; the term could also describe Ptolemy’s text. Whatever it was that Marinus was revising, Ptolemy saw himself as continuing that iterative process of revision. Ἀλλ’ εἰ μὲν ἑωρῶμεν μηδὲν ἐνδέον αὐτοῦ τῇ τελευταίᾳ συντάξει, κἂν ἀπήρκεσεν ἡμῖν ἀπὸ τούτων μόνων τῶν ὑπομνημάτων ποιεῖσθαι τὴν τῆς οἰκουμένης καταγραφὴν μηδέν τι περιεργαζομένοις· ἐπεὶ δὲ φαίνεται καὶ αὐτὸς ἐνίοις τε μὴ μετὰ καταλήψεως ἀξιοπίστου συγκατατεθειμένος καὶ ἔτι περὶ τὴν ἔφοδον τῆς καταγραφῆς πολλαχῆ μήτε τοῦ προχείρου μήτε τοῦ συμμέτρου τὴν δέουσαν πρόνοιαν πεποιημένος, εἰκότως προήχθημεν, ὅσον ᾠόμεθα δεῖν, τῇ τἀνδρὸς πραγματείᾳ συνεισενεγκεῖν ἐπὶ τὸ εὐλογώτερον καὶ εὐχρηστότερον. [3] Καὶ δὴ τοῦτο ποιήσομεν ἀπερίττως, ὡς ἔνι μάλιστα, προεπισκεψάμενοι διὰ βραχέων ἑκάτερον εἶδος τῶν ὀφειλόντων λόγου τινὸς τυχεῖν. 776 In Ptolemy’s judgement, Marinus did not correct his own work sufficiently; he was overly trusting in some instances (ἀξιοπίστου συγκατατεθειμένος) and did not sufficiently consider the theoretical aspects of map- making (τὴν ἔφοδον τῆς καταγραφῆς). As far as we can tell from his comments in book one of his Geography , Ptolemy seems to have reserved his criticism for mathematical matters, although he did make some corrections to account for the better state of knowledge of certain areas in his time. 777 774 Berggren (2000) 45–50. 775 Ptol. Geog. 1.17.1, 1.18.3. 776 Ptol. Geog. 1.6.2-3: “Now if we saw no defect in his final compilation, we would content ourselves with making the map of the oikoumene on the basis of these writings alone, without taking any more trouble with it. Since, however, even he turns out to have given assent to certain things that have not been creditably established, and in many respects not to have given due thought to the method of map-making, with a view either to convenience or to the preservation of proportionality , we have justifiably been induced to contribute as much as we think necessary to the man’s work to make it more logical and easier to use. W e will do this as concisely as possible, starting with a brief examination of each kind of thing that needs some comment.” Berggren (2000) trans. 777 He spends much of his first book correcting Marinus’ latitude and longitude calculations (1.7-14), certain 222/448 Unfortunately , Marinus’ sources are not clear from Ptolemy’s description of his work. Ptolemy notes that the coordinate data which he includes came from surveying (τὸ γεωμετρικὸν) and astronomical observation (τὸ μετεωροσκοπικόν). 778 However, Ptolemy’s criticism of his predecessor’s method of calculating the size of the oikumene shows that Marinus made use of distance information derived from movement. Ptolemy reports on a pair of distance measurements between the Euphrates and the Stone Tower and from the Stone Tower to the capital of the Seres which Marinus obtained from travellers. 779 In fact, Ptolemy includes the name which Marinus reported for the source of this information, one Maes Titianus, specifically noted to be Macedonian but also bearing an Aramaic name. 780 Thus Marinus and ultimately Ptolemy used a combination of Hellenistic science and practical route information obtained through physical movement through space. As I noted earlier, the basis of Ptolemy’s method was accumulative. He began with an already formed body of information (Marinus’ catalogue) and produced his own work by addition and correction. In this topological details (1.15-16), and his predecessor’s simplistic rectangular projection which failed to account for the curvature of the earth (1.18, 1.20; Neugebauer (1975) 879–92). He also updated the transmitted data for regions which were better known in Ptolemy’s time (1.17). 778 Ptol. Geog. 1.2. 779 Ptol. Geog. 1.11.3: “Τὴν δὲ ἀπὸ τῆς ἐκτεθειμένης τοῦ Εὐφράτου διαβάσεως μέχρι τοῦ Λιθίνου Πύργου διάστασιν συναγομένην κατ’ αὐτὸν σχοίνων ὀκτακοσίων ἑβδομήκοντα ἓξ, σταδίων δὲ δισμυρίων ἑξακισχιλίων διακοσίων ὀγδοήκοντα, καὶ τὴν ἀπὸ τοῦ Λιθίνου Πύργου μέχρι Σήρας τῆς τῶν Σηρῶν μητροπόλεως, ὁδοῦ μὲν μηνῶν ἑπτὰ, σταδίων δὲ τρισμυρίων ἑξακισχιλίων διακοσίων, ὡς ἐπὶ τοῦ αὐτοῦ παραλλήλου, συναιροῦμεν ἑκατέραν κατὰ τὴν ἐπιβάλλουσαν διόρθωσιν, ἐπειδήπερ ἐπ’ ἀμφοτέρων μὲν τῶν ὁδῶν φαίνεται μὴ μειώσας τὸ παρὰ τὰς ἐκτροπὰς πλεονάζον, ἐπὶ δὲ τῆς δευτέρας ἔτι καὶ ταῖς αὐταῖς ἀλογίαις περιπεπτωκὼς, αἷς καὶ κατὰ τὴν ὁδὸν τὴν ἀπὸ τῶν Γαραμάντων ἐπὶ τὴν Ἀγίσυμβα χώραν.” (However, we reduce according to the appropriate correction both the distance from that crossing of the Euphrates to the Stone Tower, which amounts (according to [Marinus]) to 876 schoinoi or 26,280 stades, and that from the Stone Tower to Sera, the metropolis of the Seres, a journey of seven months [ὁδοῦ μὲν μηνῶν ἑπτὰ], or [according to Marinos] 36,200 stades reckoned on the same parallel [through Rhodes]. For in the case of both journeys [τῶν ὁδῶν], [Marinos] had clearly not subtracted the excess resulting from diversions [τὰς ἐκτροπὰς], and in the case of the second, he has fallen as well into the same illogicalities that he also fell into concerning the journey [τὴν ὁδὸν] from the people of Garame to Agisymba.) Berggren (2000) trans. 780 Ptol. Geog. 1.11.5-6: “καὶ γὰρ δι’ ἐμπορίας ἀφορμὴν ἐγνώσθη. [6] Μάην γάρ φησί τινα τὸν καὶ Τιτιανὸν, ἄνδρα Μακεδόνα καὶ ἐκ πατρὸς ἔμπορον, συγγράψασθαι τὴν ἀναμέτρησιν οὐδ’ αὐτὸν ἐπελθόντα, διαπεμψάμενον δέ τινας πρὸς τοὺς Σῆρας.” (Moreover, it was because of the opportunity for commerce that [the route] came to be known. Marinus says that one Maes, also known as Titianus, a Macedonian and a merchant by family tradition, recorded the distance measurements, though he did not traverse it himself but sent certain [others] to the Seres.) Berggren (2000) trans. For discussion of this route, see Chapter 5. 223/448 method, past knowledge became anonymised layers of intellectual sediment available for future intellectual producers to build upon. This differed from authors like Strabo (and probably Marinus himself) who gathered information from various sources and constructed a new narrative by selection and substitution. The kind of sedimentation of geographical knowledge practised by Ptolemy was possible because of the slow rate of change of physical geography; for example, at the scale at which the Roman geographical writers were examining, the course of the Euphrates was a constant throughout the period. Authors like Strabo whose work was concerned at least as much with ethnography , history , and human interaction with the environment made more use of selective and substitutive techniques. Even so, the founding of cities and the movement of population groups was usually a slow process. Nicolet describes the period of geographical writing between Eratosthenes and Ptolemy as “filling in a picture whose outlines were already traced.” 781 Periods of conflict and discovery expanded the well-known and measured zone, sometimes explosively as during the period of Alexander’s conquests, but often slowly from the accounts of individuals travellers or by events of a much smaller scope, such as the greater knowledge of Dacia which followed Trajan’s campaigns. Otherwise, new information about the internal geography of the known world amounted to corrections of prior mistakes and changes in human geography . Because Marinus only survives through Ptolemy , we have no access to any geographical changes that may have occurred between the two authors. However, Ptolemy’s explicit comments about Marinus’ processes of correction and revision show that Ptolemy’s major contribution to the body of knowledge assembled by his predecessor was corrective. Something of this approach can be seen in Pliny’s treatment of his foreign geographical sources. The bulk of his work contains references to and citations of the foreign sources on which so much of his basic geographic information relied, but in the geographical chapters he overwhelmingly refers to Roman sources. In treating his sources this way , Pliny presents Hellenistic knowledge as an accumulated sediment of 781 Nicolet (1991) 63–64. 224/448 anonymised knowledge within which individual contributions need not or should not be distinguished. This focuses attention on the changes which have occurred in recent times and the corrections which that earlier material required, provided by Roman exploration, conquest and description. The Hellenistic Imprint on Mesopotamia In Chapter 1, we saw how the Roman geographical writers delimited and denominated the spaces they described. The tools they used to do so were derived from Hellenistic geographic science. For the most part, the Roman geographic writers used Hellenistic spatial categories to organise the space they described. Strabo, Pliny and Ptolemy maintained the traditional geographic divisions of Syria, Mesopotamia, Adiabene and Babylon as their Hellenistic predecessors. Only later did the writers of the fourth century adapt the categories for their own work: Ammianus Marcellinus used Roman and Persian administrative categories to delimit and denominate space, while the Expositio Totius Mundi organised space according to Roman administrative categories within that empire and according to ethnographic categories beyond it. The knowledge which Strabo, Pliny and Ptolemy drew on to describe northern Mesopotamia came primarily from the tradition of Hellenistic geographic knowledge, especially the legacy of Alexander and the Seleucid kings. The Roman imperial writers used the raw geographical material from these Hellenistic sources to populate the space of the Mesopotamian borderland with geographic and ethnographic features. The rest of this chapter will examine the interaction between the Hellenistic tradition of knowledge and the Roman construction of space in the Mesopotamian borderland. Spatial Measurement Greece and Rome had separate traditions of spatial measurement. The Roman tradition is seen in the methods of the Roman surveyors known as agrimensores. 782 These were responsible for the spatial 782 Dilke (1971) esp. 31–34. 225/448 organisation of new urban foundations and the division of agricultural land into individual allocations, known as centuriation. Although the division of land according to mathematical methods had been practised elsewhere in the Mediterranean and the Near East before Rome’s widespread foundation of colonies, centuriation was a specifically Roman practice that had much in common with the systems of spatial division practised by Roman augurs in observing bird flight including the shared use of certain technical terms. 783 Centuriation was a precise and ordered system suitable for land measurement on a local scale where sight lines could be employed and the curvature of the earth would not substantially effect the precision of measurement. It was less applicable to broad geographical description and does not appear to have influenced geographical works at that scale in a technical manner. 784 A tradition of spatial description more important for Roman geographic writing was the linear or hodological approach to space. 785 This conception of space as a series of points which could be described in a list or drawn as a line is conditioned by the experience of space as a medium of movement and travel. The scale of the hodological approach to space was more applicable to the kind of distances used in Roman geographic texts. Although this approach came to be seen widely in Roman contexts, its genesis was not specifically Roman. An itinerary is essentially a geographically ordered list. Writing as a technology is particularly conducive to conveying information in this form. 786 Many genres of Babylonian writing were composed as lists. 787 Scribes composed itineraries of the military expeditions of the Assyrian kings. 788 The Greek periplus genre described coastal voyages in a linear structure. Strabo in particular used the linear order 783 Purcell (1990) 15; Dilke (1971) 32–33, 86f. 784 Some technical concepts from centuriation were applied on a boarder scale in geographical descriptions. Livy twice uses the term cardo to demarcate mid-republican maritime space (40.18; 41.1.3), Purcell (1990) 18. The influence of Roman spatial organisation instead appears in the desire to represent space in relation and oriented towards Rome and Roman power. 785 As discussing in the Introduction. 786 Goody (1986) 36–41, 54–55; Goody (2000) 146. Goody reports that lists are one of the first forms of writing to be adopted by oral societies, Goody (1987) 211–12. 787 Beaulieu (2007). 788 Fales (1995); Liverani (1995); Kühne (1995). 226/448 provided by rivers and coastlines to structure his narrative; however, he did not use this technique in his description of Mesopotamia. 789 Pliny also structured parts of his narrative in this way , including his itinerary- like description of sites along the Euphrates in book five and his itinerary fragment from Zeugma to the Roman border in book six. 790 This hodological approach could give structure to a geographical narrative, but it could also be a source of information on the distances between places. With the exception of dedicated itineraries, linear distance measurements appear relatively rarely in ancient descriptive geographies. Where they do appear they are based either on geometric calculation or hodological travel narratives. While the source of the later simply requires movement and requires a close analysis of the context to determine the source of the knowledge, the former were produced by the close relationship between mathematics and geography in the Hellenistic period. 791 Strabo gives measurements derived from Hellenistic sources, as well as distances derived from travel. At the start of his description of Mesopotamia he provides distances between Zeugma, Thapsacus and Babylon: Ἐπὶ μῆκος δὲ συχνὸν προπέπτωκεν ἡ συναγωγὴ τῆς Μεσοποταμίας καὶ πλοίῳ πως ἔοικε· ποιεῖ δὲ τὸ πλεῖστον τῆς περιφερείας ὁ Εὐφράτης· καὶ ἔστι τὸ μὲν ἀπὸ τῆς Θαψάκου μέχρι Βαβυλῶνος, ὡς εἴρηκεν Ἐρατοσθένης, τετρακισχίλιοι καὶ ὀκτακόσιοι στάδιοι· τὸ δ’ ἀπὸ τοῦ κατὰ Κομμαγηνὴν ζεύγματος, ἥπερ ἐστὶν ἀρχὴ τῆς Μεσοποταμίας, οὐκ ἔλαττον τῶν δισχιλίων σταδίων ἕως ἐπὶ Θάψακον. 792 He credits Eratosthenes as the source of the distance measurements, but as an earlier passage makes clear, that scholar is also the source of the description of Mesopotamia as shaped like a boat (πλοίῳ πως ἔοικε). In a passage from book 2, Strabo paraphrases Eratosthenes at length including on the distance between Thapsacus and Babylon (ἀπὸ γὰρ τῆς κατὰ Θάψακόν φησι διαβάσεως παρὰ τὸν Εὐφράτην εἰς μὲν 789 Dueck (2000) 40–43. 790 Both are described in detail in Chapter 1, and discussed below at pp.199f. 791 For example see Strabo’s criticism of Posidonius as including more mathematics than proper geography in his work, 2.2.1. 792 Strabo 16.1.22: “The shape of Mesopotamia projects to a considerable length and resembles a boat. The Euphrates forms the greatest part of its periphery. From Thapsacus to Babylon, as Eratosthenes states, is four thousand eight hundred stadia; and from the Zeugma at Commagene, where Mesopotamia begins, it is not less than two thousand stadia to Thapsacus.” 227/448 Βαβυλῶνα σταδίους εἶναι τετρακισχιλίους ὀκτακοσίους), and directly quotes his metaphor for the shape of Mesopotamia (“γίνεται δή” φησί “τὸ σχῆμα τῆς Μεσοποταμίας καὶ Βαβυλωνίας ὑπηρεσίῳ παραπλήσιον.” ὁ μὲν δὴ Ἐρατοσθένης τοιαῦτ’ εἴρηκε). 793 In both of these passages, Strabo attributes to Eratosthenes the distance of 4800 stadia between Babylon and Thapsacus. In book 16, Strabo does not cite a source for his comparison of shapes, but in book 2 he also attributes to Eratosthenes the likening of Mesopotamia’s shape to that of a boat (ὑπηρεσίῳ). Strabo frequently describes the shapes of the regions he describes in such metaphorical ways. These specific images may all be indebted to Eratosthenes; certainly at least the analogic technique was. 794 As these two passages show , Strabo has kept the image of Mesopotamia’s shape as like a boat, but changed the particular word from πλοίῳ to ὑπηρεσίῳ, perhaps reflecting his particular conception of the space. Thus, at the broadest scale, Strabo’s description of Mesopotamia’s shape and dimensions is reliant on Eratosthenes, with no evidence of modification by the latter’s successors in the geographic tradition. 795 While it is unclear whether Eratosthenes’ dimensions derived from astronomical observation or travel distance calculations, they passed through the scientific tradition en route to Strabo. The distance measurements provided by Pliny are all based on travel. His description of the route down the Euphrates in book five includes distances in Roman miles between the cataracts of the Euphrates in the Taurus and Samosata, between Samosata and Zeugma, and between Zeugma and Masice in 793 Strabo 2.1.26: “For he says that from the crossing at Thapsacus along the Euphrates to Babylon is four thousand eight hundred stadia... ‘So it happens,’ he says, ‘that the shape of Mesopotamia and Babylonia is like a galley.’ Such are things Eratosthenes has said.” 794 On the shape of the oikumene as a cloak (chlamys): Strabo 2.5.6, 2.5.9, 11.11.7. Dueck (2000) 44. 795 See also Strabo 16.1.21: “τὸ μὲν οὖν μέγιστον ὃ ἀφίστανται διάστημα ἀπ’ ἀλλήλων τὸ πρὸς τοῖς ὄρεσίν ἐστι· τοῦτο δ’ ἂν εἴη τὸ αὐτὸ ὅπερ εἴρηκεν Ἐρατοσθένης, τὸ ἀπὸ Θαψάκου, καθ’ ὃ ἦν τὸ ζεῦγμα τοῦ Εὐφράτου τὸ παλαιόν, ἐπὶ τὴν τοῦ Τίγριος διάβασιν, καθ’ ἣν διέβη Ἀλέξανδρος αὐτόν, δισχιλίων τετρακοσίων· τὸ δ’ ἐλάχιστον μικρῷ πλέον τῶν διακοσίων κατὰ Σελεύκειάν που καὶ Βαβυλῶνα.” (Now the greatest distance by which the two rivers are separated is that towards the mountains; and this distance might be the same as that stated by Eratosthenes — I mean that from Thapsacus, where was the old bridge of the Euphrates, to the crossing of the Tigris, where Alexander crossed it — two thousand four hundred stadia; but the shortest distance between the two rivers is somewhere in the neighbourhood of Seleuceia and Babylon, slightly more than two hundred stadia.) 228/448 Babylonia. 796 These measurements are separated by itinerary-like lists of sites and peoples, and by an aside on Palmyra and the desert cities to the south. Taken together, these three measurements make a contiguous journey between Samosata and Masice, the first major city of Babylonia. This hints at a common source for the distances; Pliny was adept at disguising lists by splicing together multiple sources. 797 The description of Palmyra which Pliny places amidst this Euphrates itinerary also includes a number of distance measurements which connect that city with other sites of commercial significance in both Roman and Parthian space. abest ab Seleucia Parthorum, quae vocatur Ad Tigrim, CCCXXXVII [m.]p., a proximo vero Syriae litore CCIII et a Damasco XXVII propius.. 798 These distances reflect the trade routes connecting Palmyra with Babylonia, the nearby major city of Damascus and the Mediterranean shipping networks that reached out to the rest of the Roman world. The alternative method for deriving distance measurements like this would be through scientific measurement of coordinate data and geometric calculation. However, the urbanisation of the Palmyrene oasis also dates to the Roman period. Evidence for Hellenistic Palmyra is slim, but shows no signs of the widespread trade activity like that which it conducted under the Roman occupation of Syria. 799 There is little reason why such an unimportant location would be noted in coordinate charts of the Hellenistic period or used as a major marker for long distance measurements by Hellenistic mathematicians. Instead, these distances reflect the movement of long distance traders in the early empire shortly before Pliny’s time. 800 In book 6, Pliny includes a section of an itinerary leading from Zeugma to the Roman border (terminus Romani imperi) at Oruros. 801 Except for Zeugma itself, the locations in this list are obscure and 796 Pliny NH 5.85, 86, 90. On Masice, see Appendix 1.12. 797 For example, Pliny NH 5.81-82. See Chapter 1. 798 Pliny NH 5.88: “[Palmyra] is distant 337 miles from Seleucia of the Parthians, generally known as Seleucia on the Tigris, 203 from the nearest part of the Syrian coast, and 27 less from Damascus.” 799 See Chapter 5 and Appendix 1.15. 800 For a full discussion of Palmyra and Palmyrene commercial links, see Chapter 5. 801 Pliny NH 6.119-120: “Dicta est et in Zeugmate Apamea, ex qua orientem petentes excipit oppidum Caphrena munitum, quondam stadiorum LXX amplitudine et satraparum regia appellatum, quo tributa conferebantur, nunc in arcem redactum. [120] Durant, ut fuere, Thebata et, ductu Pompei Magni terminus Romani imperi, 229/448 unmentioned by other geographical sources. 802 This, along with the inclusion of a distance measurement between the two locations on the list only important to Roman space (Zeugma and the Roman border), suggest a Roman source. The explicit notice of Pompey’s border further suggests that Pliny or a secondary source found this segment in the campaign reports of a Roman general, probably those of Pompey himself. There is little reason to suspect that any of Pompey’s successors in the region would glorify his achievement of establishing a termini so far to the east of the Euphrates, especially when that river was proudly proclaimed as the border by Augustan propaganda. 803 Pliny directly quoted the acta triumphorum as the source for his discussions of aspects of Pompey’s triumph later in the Natural History. 804 Those records are a likely source for Pliny’s description of the border established by Pompey , but of the sources he lists in NH 1, any of Cornelius Nepos, M. Varro, or Livy are also possible. Delimitation and Denomination The Roman geographical writers also maintained systems of delimitation and denomination inherited from their Hellenistic predecessors. 805 The most obvious is the name Mesopotamia itself which derives from the Greek description of the space “between the rivers”. 806 Strabo, Pliny and Ptolemy all retained this name for the geographical region between the Taurus and Babylonia. Pliny’s praefectura Mesopotamiae seems to reflect his understanding of a Parthian or Assyrian categorisation. 807 After the third century establishment of an official Roman administrative category (a provincia) called Mesopotamia, the name came Oruros, a Zeugmate L·CC [m.p.].” (Apamea at Zeugma has been mentioned also, from which, heading east, one comes to Caphrena, a fortified town, formerly seventy stadia in size and called the "Palace of the Satraps." Tribute used to be brought here; now it is just a fortress. [120] Thebata is still in the same state as formerly , and Oruros, the limit of Roman power under Pompeius Magnus, 250 miles from Zeugma.) 802 Although see Chapter 1 for attempts to identify Caphrena, Thebata and Oruros. 803 Strabo 17.1.28. See Chapter 6. 804 Pliny NH 7.97-98; 37.12-19. So too did Diodorus (40.4). 805 See Chapter 1. 806 For Strabo’s definition of “Mesopotamia” as a pheronym using ἀπὸ τοῦ συμβεβηκότος (16.1.21) see Chapter 1. 807 See Chapter 1. 230/448 to signify a more restricted area. 808 Before the use of the term “Mesopotamia” to refer to a specific administrative unit, the word consistently referred to a geographical designation regardless of the political entity that nominally controlled it. Such geographical and ethnic categories generally remained constant in the geographical writers from the Hellenistic period into the Roman period. For example, the Gordyaeans and Mygdonians in Strabo and Pliny , although it is difficult to precisely know whether this was because the categories themselves remained constant or because Strabo and Pliny both drew upon the same Hellenistic sources. However, we can say that both authors selected material which resulted in at least the appearance of constancy . The main differences in the delimitation of space between Strabo and Pliny came in their representation of the imperial influence and control in the region, as will be discussed in the next chapter. In contrast to these narrative geographies, Ptolemy attempted to apply a mathematical categorisation of space over the world and largely disregarded such imperialistic perspectives. On the other side of the coin, the Expositio Totius Mundi eschewed geographical delimitation for political, administrative and ethnographic categories. Ammianus used the term “Mesopotamia” to refer to the province rather than the geographical space to which Strabo referred. Although we again have cause to lament the loss of his geographical description of Mespotamia and the knowledge of whether he treated the area so denominated as a geographical or political space. Although Ammianus had a much better understanding of the space of the Mesopotamian borderland, both through the gradual accumulation of available knowledge of a now imperial province and through his detailed personal experience of the area, some conceptions of the historical geography of the Mesopotamian borderland remained from the beginning of the Roman period. The conception of the Macedonian conquest as a civilising movement persisted. Ammianus made much the same claim as Pliny in his description of the eastern frontier regions. 808 See Chapter 2. 231/448 orientis vero limes in longum protentus et rectum ab Euphratis fluminis ripis ad usque supercilia porrigitur Nili, laeva Saracenis conterminans gentibus, dextra pelagi fragoribus patens, quam plagam Nicator Seleucus occupatam auxit magnum in modum, cum post Alexandri Macedonis obitum successorio iure teneret regna Persidis, efficaciae inpetrabilis rex, ut indicat cognomentum. [6] Abusus enim multitudine hominum, quam tranquillis in rebus diutius rexit, ex agrestibus habitaculis urbes construxit multis opibus firmas et viribus, quarum ad praesens pleraeque licet Graecis nominibus appellentur, quae isdem ad arbitrium inposita sunt conditoris, primigenia tamen nomina non amittunt, quae eis Assyria lingua institutores veteres indiderunt. 809 In this appeal to the Macedonian development of now-Roman provinces, Ammianus’ account strongly resembles the description of Mesopotamia in Pliny’s sixth book. 810 Like Pliny , Ammianus begins the history of the space with the Macedonian conquest and Seleucus’ inheritance and development of the region. The Achaemenid Persians appear only as Alexander’s defeated opponents. Ammianus casts the Macedonian conquest as a positive influence on the region (plagam Nicator Seleucus occupatam auxit magnum), implicitly equating wealth with urbanisation (ex agrestibus habitaculis urbes construxit multis opibus firmas et viribus). This equation fits with the Roman ideal of provincial organisation. Geographical Material The geographic writers of the early empire were dependent on Hellenistic material as well as methodology in their geographic narratives. This can be seen in the sheer quantity of Hellenistic information that the works of Strabo, Pliny and Ptolemy contained, most obtained from literary sources. None of those authors made grandiose claims of widespread autopsy . Strabo comes closest when he claims to 809 Amm. Marc. 14.8.5-6: “But the frontier of the East, extending a long distance in a straight line, reaches from the banks of the Euphrates to the borders of the Nile, bounded on the left by the Saracen peoples and open on the right to the waves of the sea. Seleucus Nicator took possession of the region and greatly increased it in power, when, after the death of Alexander of Macedon, he was holding the rule of Persia by right of succession; a capable and successful king, as his surname [Nicator] indicates. [6] For he took advantage of the great number of men whom he ruled for along time in peace, and built cities that were secure in wealth and power out of their rustic dwellings. Now most of these are called by the Greek names which their founder imposed upon them, nevertheless have not lost the original names which their ancient settlers gave them in the Assyrian language.” 810 Pliny NH 6.117 (discussed above at p.197): “Mesopotamia tota Assyriorum fuit, vicatim dispersa praeter Babylona et Ninum. Macedones eam in urbes congregavere propter ubertatem soli”(All Mesopotamia belonged to the Assyrians, the population scattered in villages except for Babyon and Ninevah. The Macedonians gathered them together in cities because of the fertility of the soil.) 232/448 have travelled further than any other single geographer. 811 However, he immediately follows this claim with an argument for the validity of relying on second hand geographical description. Strabo places literary sources on the same level as sense perception for the accumulation of geographical knowledge. Pliny goes further and prioritises secondary research to personal observation, while Ptolemy likewise makes no claim to autopsy and places himself in the position of being a corrector of a longer tradition, and Marinius in particular. All three authors thus worked within a framework which prioritised the transmission of knowledge between and through existing texts rather than the collection and creation of new knowledge. The novelty in their work derived from selection, arrangement, and presentation. Structural and methodological factors mark their work as Hellenistic as we have seen, but so to does the material they selected to represent the space. The latter applies particularly to Strabo and Pliny . Ptolemy displays a lower degree of selectivity , instead aiming at a more exhaustive treatment which built upon the body of evidence that his predecessor had collected and focused on the correction of factual and technical errors. The space that he constructs is Hellenistic in that it is recorded and transmitted within the sedimentary structures of Hellenistic science. The material handed down by Ptolemy attests to the collection of coordinate data by his predecessors in the Hellenistic tradition of mathematical geography . Strabo and Pliny also reflect and construct Hellenistic space through these structures, but to a larger degree, they do so through their geographic material. The inclusion of Macedonian spaces, places and events locates Macedonian history in the geographical space of Mesopotamia and locates that space within the stage of Macedonian history . The geography of Mesopotamia is thus a imperialist colonial history . 812 A brief survey of the accounts of Strabo and Pliny will show the extent to which their material constructs Mesopotamia as a Macedonian colonial space. Strabo’s description of Mesopotamia is filled with references to Hellenistic conquests, scholarship and 811 Strabo 2.5.11. Quoted and discussed above, p.197. 812 Purcell (1990). 233/448 colonisation. The first two sections (16.1.21-22) refer to Alexander’s crossing of the Tigris before the battle of Gaugamela, mention Eratosthenes three times, and use the important Macedonian foundations of Zeugma and Seleucia on the Tigris as reference points for long-distance measurements. His description of the fertile area near the mountains (16.1.23) names eight locations, of which two were Macedonian foundations with Greek names (Zeugma and Nikephorion), two were Macedonian colonies founded in major cities which retained their local names (Nisibis and Carrhae), one was a Hellenistic foundation of the Hellenised Pontic kingdom (Tigranocerta), one has an Iranian name (Chordiraza), and two had Aramaic names (Sinnaca and Thapsacus). 813 This apparent diversity is reduced if the eight sites are expressed differently: four were apparent as Macedonian colonies either through their relative fame or the Hellenic nature of their name (Nisibis, Carrhae, Zeugma and Nikephorion), two more were important sites in the Roman history of the region (Tigranocerta and Sinnaca, as well, of course, as Carrhae), and one would have been well-known by a reader familiar with the campaigns of Xenophon and Alexander (Thapsacus). 814 Sites with Greek allusions predominate. To these should be added the sites which Strabo located in this northern area in his description of the commercial route down the Euphrates: Bambyce/Hierapolis/Edessa and Anthemusia. 815 The three sites reflected in these names bear respectively a Greek name, and the names of two Macedonian cities. 816 Turning to Pliny , we began this chapter with his explicit contrast between the humble, dispersed population of Mesopotamia under the Assyrians and the prosperous, concentrated population under the Macedonians. 817 Having been primed for the Macedonian transformation of the region, the reader can clearly see the imprint of the Macedonian presence in the Greek and Macedonian dynastic names that follow . Oppida praeter iam dicta habet Seleuciam, Laodiceam, Artemitam; item in Arabum gente qui 813 Strabo 16.1.23. For all of these sites, see Chapter 1 and Appendix 1. 814 Leaving the enigmatic Chordiraza, Appendix 1.7. 815 Strabo 16.1.28. For a discussion of this route, see Chapter 5. 816 For an examination of Strabo’s equation of Bambyce/Hierapolis and Edessa, see Chapter 1. For all three sites, see Appendix 1. 817 Pliny , NH 6.117, quoted and translated on p.197. 234/448 Orroei vocantur et Mandani Antiochiam quae a praefecto Mesopotamiae Nicanore condita Arabis 818 vocatur. [118] Iunguntur his Arabes introrsus Eldamari, supra quos ad Pallacontam flumen Bura oppidum, Salmani et Masei Arabes; Gurdiaeis vero iuncti Azoni, per quos Zerbis fluvius in Tigrim eadit, Azonis Silices montani et Orontes, quorum ad occidentem oppidum Gaugamela, item Suae in rupibus. Supra Silicas Sitrae, per quos Lycus ex Armenia fertur, ab Sitris ad hibnernum exortum Azochis oppidum, mox in campestribus oppida Dios Pege, Polytelia, Stratonicea, Anthemus. [119] In vicinia Euphratis Nicephorion, quod diximus; Alexander iussit condi propter loci opportunitatem. Dicta est et in Zeugmate Apamea; ex qua orientem petentes excipit oppidum Caphrena munitum, quondam stadiorum LXX amplitudine et Satraparum Regia appellatum quo tributa conferebantur, nunc in arcem redactum. [120] durant, ut fuere, Thebata et, ductu Pompei Magni terminus Romani imperi, Oruros, a Zeugmate CCL. 819 In addition to the cities which simply bear witness to Macedonian rule through their names (such as Seleucia, Laodicea, and Apamea), Pliny singles out for specific comment the foundations of “Arabian” Antiochia and Nicephorion; the former by Nicanor (“a praefecto Mesopotamiae Nicanore condita”) and the latter by Alexander (“Alexander iussit condi propter loci opportunitatem”). Two of the most important sites in Mesopotamia retained their local names despite the presence of Macedonian colonists. The main city of the Mygdonians was Nisibis, a city refounded by a Seleucid as Antiocheia in Mygdonia. 820 This important local centre had a Macedonian name which never appears as the primary toponym of the city . Here, as with many Antiochs, the Macedonian name did not stick, either because of the strength of local traditions or because of the need to differentiate among a superfluity of 818 Arabis (or Arabes) is a conjecture for “Arabs” in the manuscript tradition. 819 Pliny , NH 6.117-120: “Except for the towns already mentioned, [Mesopotamia] contains Seleucia, Laodicea, Artemita; and in Arabia, the people called the Orroei, and Antiochia of the Mardani, founded by Nicanor, the governor of Mesopotamia, and called Arabis*. [118] Joined to these in the interior are the Eldamari Arabs (beyond whom is the town of Bura near the river Pallaconta) and the Salmani and Masei Arabs. Next to the Gordyaei are the Azoni, through whose territory the river Zerbis runs into the Tigris; next to the Azoni are the Silici, a mountain tribe, and the Orontes, to the west of whom lies the town of Gaugamela, as also Suae among the rocks. Beyond the Silici are the Sitrae through whose district the river Lycus flows out of Armenia, south- east of Sitrae the town of Azochis, then in the plains the towns of Dios Pege (Zeus’ Spring), Polytelia, Stratonice, and Anthemus. [119] In the vicinity of the Euphrates is Nicephorion, which we have mentioned; Alexander ordered it to be built because of the favourableness of the site. Apamea at Zeugma has been mentioned also, from which, heading east, one comes to Caphrena, a fortified town, formerly seventy stadia in size and called the "Palace of the Satraps." Tribute used to be brought here; now it is just a fortress. [120] Thebata is still in the same state as formerly , and Oruros, the limit of Roman power under Pompeius Magnus, 250 miles from Zeugma.” 820 Strabo 16.1.23: “...ἡ Νίσιβις, ἣν καὶ αὐτὴν Ἀντιόχειαν τὴν ἐν τῇ Μυγδονίᾳ προσηγόρευσαν...”. See also Strabo’s initial survey of the fertile crescent (16.1.1) where he calls the Mygdonians “τοὺς περὶ Νίσιβιν Μυγδόνας”. For Nisibis and its the foundation by Nicanor, see Appendix 1.14. 235/448 Antiochs. Nevertheless, both Strabo and Pliny report that name. 821 To the east, Carrhae apparently received colonists in the Macedonian period. These either retained, or were plausibly able to claim, sufficient numbers and philhellenism to support the legions of Afranius and Crassus when they passed through the area in 65 and 54-53 BCE respectively . 822 Both Strabo and Pliny mention the town, but not the Hellenistic colonists; instead they recall Crassus’ defeat. 823 Finally , it should be remembered that most of the cities of northern Mesopotamia lay within the area which both authors call Mygdonia. This name, the same as a river in Macedonia and related to a Homeric hero, was most likely an interpretatio Graecae for the Semitic name of some local political, social or cultural group. 824 Pliny is explicit on the Macedonian renaming of the space: Namque Persarum regna, quae nunc Parthorum intellegimus, inter duo maria Persicum et Hyrcanium Caucasi iugis attolluntur. utrimque per devexa laterum Armeniae Maiori a frontis parte, quae vergit in Commagenen, Cephenia, ut diximus, copulatur eique Adiabene, Assyriorum initium, cuius pars est Arbilitis, ubi Darium Alexander debellavit, proxime Syriae. [42] totam eam Macedones Mygdoniam cognominaverunt a similitudine. oppida Alexandria, item Antiochia quam Nesebin vocant; abest ab Artaxatis DCCL [m.]p. fuit et Ninos, inposita Tigri, ad solis occasum spectans, quondam clarissima. 825 Pliny is imprecise on the exact bounds of the area which the Macedonians named Mygdonia. By including 821 Strabo 16.1.23; Pliny NH 6.42. But see Pliny NH 6.117 where “Mardani Antiochiam” might be a distorted form of Mygdonian Antioch. 822 Dio 37.5.5: (Afranius, returning through Mesopotamia to Syria, contrary to the agreement made with the Parthian, wandered from the way and encountered many hardships by reason of the winter and the lack of supplies. Indeed his troops would have perished, had not the Carrhaeans, Macedonian colonists who dwelt somewhere in that vicinity , received him and helped him forward.) Loeb trans. Dio 40.13.1. See Arnaud (1986). 823 Pliny NH 5.86: “Carrhas, Crassi clade nobile.” Although Strabo does not mention Crassus in direct apposition with Carrhae, but rather with Sinnaca shortly after. Strabo 16.1.23: “τὰ περὶ Κάρρας καὶ Νικηφόριον χωρία καὶ Χορδίραζα καὶ Σίννακα, ἐν ᾗ Κράσσος διεφθάρη, δόλῳ ληφθεὶς ὑπὸ Σουρήνα τοῦ τῶν Παρθυαίων στρατηγοῦ.” 824 See Chapter 1. 825 Pliny NH 6.41-42: “The kingdom of the Persians, which we now know as that of the Parthians, was established on the heights of the Caucasus mountains, between two seas, the Persian and Hyrcanian. As I said, Cephenia is joined to Greater Armenian on both sides by steep slopes running towards the front part which faces Commagene, and to this, Adiabene, where the land of the Assyrians begins, of which the part closest to Syria is Arbelitis, where Alexander defeated Darius. The Macedonians have named this all Mygdonia from its similarity [to Mygdonia in Macedon]. Its towns are Alexandria and Antiochia, which they call Nesebis; it is 750 m.p. from Artaxata. Ninos (Nineveh), placed on the Tigris looking to the west was once very famous.” 236/448 Adiabene, he certainly referred to a slightly different area than Strabo, but the Macedonian focus is clear in both the naming of the region and the major cities he includes within it: Alexandria and Antioch “quam Nesebin vocant”. 826 Alexandria of Mygdonia is thought to lie near Arbela. It was probably founded in 331 BCE to commemorate the Macedonian’s famous victory at nearby Gaugamela, as Pliny’s note on Arbelitis recalls (“Arbilitis, ubi Darium Alexander debellavit”). 827 Conclusion One legacy of Alexander’s Asian campaigns was the wide distribution of Macedonian and Greek colonies throughout the space occupied by his ephemeral empire. The consolidation of this space into Hellenistic kingdoms through the continuing processes of Macedonian colonisation and the development of royal control spread Hellenic culture widely , but also provided the conditions for the concentration of the new forms and systems of knowledge in royal centres like the library at Alexandria. The origin of Hellenistic culture in Greece and Macedonia meant that the Mediterranean remained the focus, at least partially , of many of the Hellenistic kingdoms and drew much of this Hellenistic knowledge into what would become the Roman sphere. 828 The geographical writers of the Roman Empire inherited the systems of organising and constructing knowledge from their Hellenistic predecessors, as well as the raw geographical material organised and created by the Macedonian colonial project. Although the Roman authors constructed geographies centred on Rome and Roman power, this Hellenistic material formed the basis of their geographic works. Strabo and Ptolemy consciously placed themselves in and appealed to the intellectual authority of the Hellenistic 826 For the delimitation of these areas, see Chapter 1. 827 Dillemann (1962) 160, n.3; ‘Alexandrian Foundations (1)’ PECS, p.39; Cassius Dio mentions the location in his description of Trajan’s invasion: Dio 68.26; Theophylactus (5.7.10-11) calls it Alexandriana. 828 In the case of the Seleucid Empire, the cultural pull of Macedonia, conflicts with western successor states like Ptolemaic Egypt, and military failures on the Iranian plateau meant that Antioch on the Orontes remained an important site of royal activity even when Seleucia on the Tigris was a more important centre for the empire as a whole. Sherwin-White and Kuhrt (1993). 237/448 tradition through specific references to Greek geographers like Eratosthenes, Posidonius, Hipparchus and Marinus. Pliny prioritised Roman sources of knowledge but his geography of Mesopotamia reflects a foundation in Hellenistic intellectual production. Pliny explicitly claims his authority through his use of Roman sources, but, at least in his descriptions of the Roman Near East, his material constantly calls to mind the Hellenistic basis of his work. Other local and foreign sources of knowledge can be occasionally seen in these texts, but are largely subsumed within the Hellenistic intellectual project. 829 Nevertheless, these authors used and adapted Hellenistic knowledge to create a Roman perspective on the Mesopotamian borderland. To do so, they selectively drew on these sources. Strabo chose his sources carefully , engaging only with those he considered the highest quality , such as Eratosthenes and Poseidonius. A comparison between the two versions of the drowning of Tryphon’s army , one directly cited from Poseidonius by Athenaios, the other clearly derived from Poseidonius, but not credited by Strabo, shows that Strabo was willing and able to use and rearrange the best sources to construct a narrative emphasising elements of his choice. Pliny’s smooth and consistent construction of Syrian geography without reference to any of his Greek sources shows his skill at selection, aggregation and arrangement of a consistent narrative. In the sphere of geography , as elsewhere, Hellenistic knowledge became an authority to be drawn on selectively in order to constitute and reinforce a Roman perspective on space, history , and imperial power, as I will show in the next chapter. 829 As we will see in Chapter 4, Strabo makes different use of that Assyrian past, but he highlights the Macedonians influence on the landscape of Assyria by frequent reference to Alexander. 238/448 Chapter 4: The Representation of Imperial Power In the previous chapter, we saw how the geographical writers of the early Roman empire relied on Hellenistic sources and techniques for the materials and tools they used to construct the space. I argued that these geographical writers did more than just mine the work of their Hellenistic predecessors for data and present a Hellenistic image of the space. Rather, they played an active part in the translation of this material and, in doing so, produced a Roman space from that material. In the sphere of geography , as elsewhere, Hellenistic knowledge became an authority to be drawn on selectively in order to constitute and reinforce a Roman perspective on space, history , and imperial power. 830 As Rome expanded into the Near East, her power was checked by the Iranian empire of the Arsacid dynasty , known to Greek and Roman sources as the Parthians. In the territory of the former Seleucid kingdom, one of the major Hellenistic kingdom’s founded by Alexander’s successors, Roman and Parthian territory alike lay atop a deep base of previous Near Eastern imperial histories. By the first century CE, the western edge of the Mesopotamian borderland was administered, directly or indirectly , by the Roman governor of Syria; the eastern edge was under the political control of the Parthians, often indirectly . This inter-imperial borderland forced Roman writers to negotiate an ideology of Roman power that acknowledged no limit (imperium sine fine) 831 in the face of a real foreign power that successfully resisted the Roman advance. This chapter examines how the geographical writers of the Roman empire represented those overlapping historical and contemporary imperial histories in their works and how they described and distinguished the two facing areas of imperial control. The chapter begins with three approaches to Strabo’s text in which I argue that the manner in which Strabo describes Semiramis, Alexander and the Macedonians 830 Gruen (1990) 79–123; Habinek (1998). 831 Vir. Aen. 1.279. 239/448 and the Parthians themselves all minimise the historical and contemporary significance of Parthian rule in the region. The second part examines how differences in the narrative style of Pliny’s descriptions of the Mesopotamian borderland in books five and six construct a dichotomy between Roman order and Parthian disorder. Finally , the chapter concludes by placing these strategies in the context of later Roman geographies generally and other writers on the Mesopotamian borderland specifically . Strabo Drawing Borders Semiramis Semiramis appears at the beginning of Strabo’s description of Assyria in book 16. While this appearance is brief, it is significant. In this section, I will discuss the appearance of Semiramis in Greek literature up to the Augustan period, then examine how Strabo used her in the Geography. I will argue that Strabo establishes Semiramis as a foundational figure in the Mesopotamian borderland and in doing so, emphasises the greatness of Mesopotamia’s distant past compared to the somewhat cursory treatment of Parthian power in the present. He represents the region as an ambiguous borderland between Roman and Iranian space; beyond Roman control but not beyond Roman power. At the start of book 16, Strabo’s Geography defines the space that will be the focus of the first half of that book using the identifier οἱ Ἀσσύριοι, the Assyrians. As the subsequent lines explain, the land of the Assyrians encompassed the lowland Asian parts of the fertile crescent from the head of the Persian Gulf to the head of the Red Sea. 832 Strabo proceeds to define who the Assyrians were: οἱ δ’ ἱστοροῦντες τὴν Σύρων ἀρχὴν ὅταν φῶσι Μήδους μὲν ὑπὸ Περσῶν καταλυθῆναι Σύρους δὲ ὑπὸ Μήδων, οὐκ ἄλλους τινὰς τοὺς Σύρους λέγουσιν ἀλλὰ τοὺς ἐν Βαβυλῶνι καὶ Νίνῳ κατεσκευασμένους τὸ βασίλειον· ὧν ὁ μὲν Νίνος ἦν ὁ τὴν Νίνον ἐν τῇ Ἀτουρίᾳ κτίσας, ἡ δὲ τούτου γυνή, ἥπερ καὶ διεδέξατο τὸν ἄνδρα, Σεμίραμις, ἧς ἐστι κτίσμα ἡ Βαβυλών. οὗτοι δὲ ἐκράτησαν τῆς Ἀσίας, καὶ τῆς Σεμιράμιδος χωρὶς τῶν ἐν Βαβυλῶνι ἔργων πολλὰ καὶ ἄλλα κατὰ πᾶσαν γῆν σχεδὸν δείκνυται ὅση τῆς ἠπείρου ταύτης ἐστί, τά τε χώματα ἃ δὴ καλοῦσι 832 See Chapter 1. 240/448 Σεμιράμιδος, καὶ τείχη καὶ ἐρυμάτων κατασκευαὶ καὶ συρίγγων τῶν ἐν αὐτοῖς καὶ ὑδρείων καὶ κλιμάκων καὶ διωρύγων ἐν ποταμοῖς καὶ λίμναις καὶ ὁδῶν καὶ γεφυρῶν. ἀπέλιπον δὲ τοῖς μεθ’ ἑαυτοὺς τὴν ἀρχὴν μέχρι τῆς Σαρδαναπάλλου καὶ Ἀρβάκου· μετέστη δ’ εἰς Μήδους ὕστερον. 833 Although Strabo calls the inhabitants of the region Syrians (Σύροι) rather than Assyrians (Ἀσσύριοι), there is no doubt he is talking about the same people. 834 The terms Syria and Assyria were used somewhat interchangeably by classical authors. 835 Using a recently published Hieroglyphic Luwian and Phoenician bilingual inscription, Robert Rollinger has demonstrated that the two terms were also used interchangeably in 8 th century BCE Anatolia and were most likely encountered by the Greeks as synonyms already . 836 Later, Herodotus equates Syria and Assyria and defines the latter as the lands inhabited by the Assyrians. 837 In this passage Strabo conflates the works of multiple empires into the reigns of one royal couple, in particular the Neo-Assyrian empire which ruled from the area around Ninos/Nineveh and the various empires centred on Babylon. He assigns the foundation of those cities to Ninus and Semiramis respectively . When he used Semiramis in his geographical narrative, Strabo was participating in a long tradition of manipulation and adaptation. The figure of Semiramis was based on two Neo-Assyrian queens who seem to have originated from the Syrian coast: Sammuramat, who lived in the late-ninth century BCE, and 833 Strabo 16.1.2: “When those who have written about the empire of the Syrians say that the Medes were destroyed by the Persians and the Syrians by the Medes, they mean by the Syrians none other than those who constructed the palaces in Babylon and Ninus [i.e. Nineveh], of whom Ninus founded Ninus in Assyria, and his wife, Semiramis, who succeeded her husband, is the founder of Babylon. They ruled over Asia, and, apart from those in Babylon, almost the entire land shows many other works of Semiramis, as many as there are in the land: the banks known as those of Semiramis, and the walls, and the construction of defences with pipes inside them, and of reservoirs, and ladders, and of canals in rivers and lakes, and of roads and dams [or bridges]. They left these things with their empire until Sardanapalus and Arbaces, and later it passed to the Medes.” The penultimate clause is corrupt, but the sense is clear, Radt (2002) 4.276; Radt (2002) 8.253. 834 Strabo considered Assyria to be a unitary area inhabited by “Syrians” who “extend from Babylonia to the Gulf of Issus” (16.1.2; 2.1.31). Strabo defines Assyria as including the Roman province of Syria in theory , but then treats it differently in practice. 835 Rollinger (2006). Hdt 7.63. Frye (1992) 283; contra Parpola (2004) 21; and less convincingly Joseph (1997). 836 Rollinger supports the main thesis of Frye (1992); Parpola also notes that unstressed initial vowels or syllables were often dropped in neo-Assyrian, (2004) 16–18. 837 Hdt. 2.17: “...Αἴγυπτον μὲν πᾶσαν εἶναι ταύτην τὴν ὑπ’ Αἰγυπτίων οἰκεομένην, κατά περ Κιλικίην τὴν ὑπὸ Κιλίκων καὶ Ἀσσυρίην τὴν ὑπὸ Ἀσσυρίων...” (...Egypt is all that country which is inhabited by Egyptians, even as Cilicia and Assyria are the countries inhabited by Cilicians and Assyrians severally...) Loeb trans; 7.63: “Οὗτοι δὲ ὑπὸ μὲν Ἑλλήνων ἐκαλέοντο Σύριοι, ὑπὸ δὲ τῶν βαρβάρων Ἀσσύριοι ἐκλήθησαν.” (These are called by Greeks Syrians, but the foreigners called them Assyrians.) Loeb trans. 241/448 Naqi’a, who lived in the eighth and seventh centuries BCE. 838 In the fifth century , these queens both appear in Herodotus as queens of Babylon with active building programs. 839 The legends surrounding these queens were elaborated and focused on the figure of Semiramis by Ctesias of Cnidus in his Persica at the beginning of the 4 th century BCE. As a physician working in the court of Artaxerxes II (404-359 BCE), Ctesias drew on oral and written sources of history and legend circulating at the Persian court to write his 23 book Persica. 840 Judging by the fragments preserved by Diodorus Sicilus in his Historical Library (first century BCE), Semiramis played a central role in Ctesias’ history of Assyria which comprised the first three books of his Persica. 841 Ctesias was a widely read author and may have been responsible for transmitting the coherent image of Semiramis which came to inhabit the Greek imagination from the late fifth century onward. Ctesias’s elaboration and transmission of the Semiramis legend led to a diverse range of adaptations of that legend for various narrative purposes. Ctesias himself used the figure of Semiramis to reflect on the 838 Sammuramat, wife of Šamšī-Adad V (r. 824-811 BCE) and appeared as queen mother in the first years of her son’s reign, Adad-Nirārī III (r. 811-783 BCE), Frahm, “Semiramis’ BNP . Naqi’a (Zakutu in Akkadian; Nitocris in Greek), wife of Sennacherib (r. 705-681 BCE) and influential in the courts of her son Esarhaddon (r. 681-669 BCE) and grandson Assurbanipal (r. 669-627 BCE), Melville (1999); Lewy (1952); Röllig, “Nitocris” BNP . The figure of Ninus was also created from a blend of Šamšī-Adad V (husband of Sammuramat) and Sennacherib (husband of Nitocris). Both Sammuramat and Naqi’a seem to have been of Syrian origin, W einfeld (1991); Lewy (1952). 839 Semiramis built dykes (χώματα) to restrict the flow of the Euphrates and prevent flooding (Hdt. 1.184). Nitocris altered the flow of the Euphrates to protect Babylonia from the Medes (Hdt. 1.185), built an elaborate bridge to allow access between the two sides of Babylon (Hdt 1.186), and built her tomb above what Herodotus calls the most important gate of Babylon (Hdt 1.187). Some sign of the blending of the works of these two queens can already be seen in Herodotus. At 3.155, he lists several gates of Babylon, including a Gate of Semiramis, but no Gate of Nitocris. 840 FGrH 688 F 1b = Diod. 2.1.4-28.7. Semiramis occupies 2.1.4-20.3. At 2.20.3, Diodorus credits Ctesias with the details of Semiramis’ life. On Ctesias’ sources, see Llewellyn-Jones and Robson (2010) 55–65. On the role of orality in the Babylonian intellectual tradition, see Beaulieu (2007). 841 Ctesias may have been responsible for introducing the divine elements to the Semiramis legends. In Diodorus’s account of Semiramis’ origins, she is born in Ascalon on the Syrian coast to the goddess Derceto, whom Strabo and Pliny equated with Atargatis: Diod. 2.4.1-6; Strabo 16.4.27; Pliny 5.81. Lucian (De Dea Syria 14) reports the same story of Semiramis’ divine birth, but claims that Atargatis and Derceto and were different, but see Oden (1977) 69–70. F 1c = Anonymous, On W omen, 1 (Llewellyn-Jones and Robson, p.139 ); F 1m = Athenagoras, Embassy for the Christians, 30 (L-J and R, p.142). On the W estern Syrian links between Atargatis, Derceto and Semiramis, see W einfeld (1991). In particular, he argues that the Greek name Derceto is derived from the Ugaritic word darkatu (meaning dominion) which appears as a epithet of Atargatis alongside šamīm ramīm, “mistress of the high heavens”. The two epithets are collocated in the Ugaratic text RS 24.252 (Schaeffer et al. (1939) V .551 Text 2.6). For divine links between Semiramis and Iranian dieties at Behistun, see Phillips (1972). 242/448 nature of power in the court of an absolute monarch by placing her in comparison to the Persian queen Parysatis, mother of Artaxerxes II and Cyrus the Younger and instrumental in the Achaemenid palace politics which surrounded Cyrus’ invasion of Babylonia in 401 BCE. 842 Diodorus actively summarised, selected and reworked his sources to derive his own conception of historical reality . 843 Diodorus used Semiramis to highlight aspects of Alexander’s campaigns and as an example of good rulership. 844 At about the same time as Diodorus, Pompeius Trogus (as transmitted by Justin) used Semiramis to ellucidate themes of monarchy , succession, and empire. 845 Semiramis’s romantic and familial relationships were also used for various purposes. 846 By the time Strabo was writing at the beginning of the 1 st century CE, Semiramis and her story had become both well-known and traditional; it could be alluded to, directly referenced and understood without full explanation. Because of this expectation of audience familiarity , Semiramis was also available for adaptation and reuse to serve a wide range of narrative purposes and didactic ends. Strabo could have started his description of Assyria with any number of important Assyrian, Babylonian, Seleucid, Achaemenid or Arsacid kings. He chose Semiramis, placing her in this prominent position as a deliberate move to contrast her role in the region with that of the Parthians. Semiramis appears several times in Strabo’s Geography , most extensively at the start of book 16. 847 In that passage, Semiramis appears as a conquerer alongside her husband Ninus, but more extensively as a 842 Llewellyn-Jones (2010) 76, 84–86. Xen. Anab. 1.1. 843 Compoli (2000); Llewellyn-Jones (2010) 39, 84. For example, the only verbatim fragment of Ctesias (Papyrus Oxyrhynchus 2330 = FGrH 688 F 8b, text and commentary in Lobel and Roberts (1954) 81–84. Bigwood (1986); Stronk (2008)) differs in important ways from the version transmitted by Diodorus (Diod. 2.34.3-5 = FGrH 688 F 5 §34.3-5; Llewellyn-Jones (2010) 37, 39–40). 844 Comparison to Alexander’s campaigns: Sulimani (2005). Moral example: Llewellyn-Jones (2010) 39. 845 Justin 1.2. Yardley (1994) 7. Yardley (2003) 846 Ctesias used a love triangle motif to embed romantic (and ultimately tragic) conflict within his novelistic treatment of the Semiramis story (Llewellyn-Jones and Robson (2010) 71); Justin used Semiramis’ disguised gender to contrast her active masculine rule with the decadent and effeminate latter Assyrian kings (Justin 1.2.3- 6; cf. Diod. 2.6.6; 2.21.2-7; 2.23.1); the Ninus Romance (Trajanic date) show that novelist contrasting the courtship behaviours of Ninus and Semiramis to show a model of the appropriate expression of feminine desire (Anderson (2009)). 847 Strabo Geog. 16.1.2 (quoted above, p.273); 2.1.16; 2.1.31; 15.1.5-6; 15.2.5. 243/448 builder of several notable Babylonian structures (fortifications, water systems and transportation works). This brief but relatively detailed reference to Semiramis is important for three reasons. First, it is one of the few occasions on which Strabo follows a broad regional overview with a historical episode from the pre- Hellenistic past. Second, it dwells specifically and personally in that bounded pre-Hellenistic past. Third, it prioritises the energetic building activity of Semiramis over her conquests. These three factors emphasise the antiquity of the region and prioritise that antiquity almost to the exclusion of present political conditions. Throughout the 17 books of Strabo’s Geography , several narrative patterns are evident. His work proceeds regularly around the Mediterranean, often dividing space according to broad provincial or ethnic units. Within each of these, he begins by defining the space to be described, its boundaries and any internal subdivisions. He then proceeds around those subdivisions, defining the space and then describing further internal subdivisions. Although historical matters appear frequently in his work, they are usually hooked into the narrative according to their relationship with some geographical feature. 848 He seldom includes broader historical overviews and these rarely stand in a prominent place at the beginning of the relevant narrative. 849 Assyria is the exception. Here, Strabo begins with a historical overview of pre-Hellenistic history focused on Semiramis and the Assyrians. By choosing to highlight the history of the region, Strabo highlights his own selective processes. In book 16, Strabo is about to treat the main region of contact between Parthia and Rome, Parthia and the Seleucids, and Alexander and Persia. Instead of giving the prominent beginning place to one of these historical episodes he looks back further to ancient history . 848 For example: 16.2.4 where the foundations of several Seleucid kings are mentioned; 16.2.8 where Strabo refers to the locations of conflicts between V entidius and the Parthians Pacorus and Phranicates, between Ptolemy Philometor and Alexander Balas, and between Pompey and Tigranes. 849 Regions described in the first instance by shape, descriptions and divisions (often with an ethnographic element): Iberia (Strabo Geog. 3.3), Gaul (4.1), Italy (5.1), Northern Europe (7.1), Asia generally (11.1), Scythians (11.2), Cappadocia (12.1), the southern coast of Anatolia (14.1), Ariana (15.2), Persis (15.3), the subdivisions of Syria and Arabia (16.2, 16.3, 16.4). Regions for which the first sections are methodological or historiographical: the Troad (13.1), Asia (15.1) as a prelude to India (which begins at 15.1.11), Egypt (17.1). Strabo’s introduction to Greece (8.1) begins with a historiography then a linguistic ethnography. Assyria (16.1) is the only major area to start with a historical outline. 244/448 Having immediately drawn the reader’s attention to the distant past, Strabo brings that past into sharp definition with the specificity of his description of Semiramis’ deeds as a builder. 850 Strabo gives her credit for not just the foundation and construction of Babylon, but for a wide range of engineering works throughout Asia. The rest of his description of Southern Mesopotamia refers to such works repeatedly , especially Babylon itself and the canal system of Babylonia. 851 These works are seldom specifically tied to Semiramis, but they are seldom attributed at all, except when Alexander of Macedon is involved. 852 Thus, primed by the introduction, the occasional reminders of past construction serve to recall the Assyrian past and remind the reader of Semiramis’ role in the region. 853 While this introduction credits Semiramis by name for her works, similar foundations by Parthian rulers are left anonymous. His arrangement of the chapter on Assyria emphasises the importance of Babylon and Seleucia. In his description of Babylon, Strabo notes that the formerly great city has since been eclipsed. 854 He names Babylon’s successor as the important Macedonian foundation of Seleucia, but it is only much later in the book that he describes the contemporary situation in which the Parthians had turned Ctesiphon from a village into an important political centre. 855 Strabo does not credit any specific Parthian king with this foundation, but rather vaguely attributes it to them collectively and impersonally . Compared to the energy of Semiramis’ foundation of Babylon, Ctesiphon has the sense of being a happy accident. Strabo’ technique of placing Semiramis’ building achievements in the foreground and highlighting them contrasts with his treatment of her conquests elsewhere in his work. In his introduction to book 16, 850 Strabo’s introduction to Assyria credits both Ninus and Semiramis as important builders of Nineveh and Babylon respectively , but the former city is quickly disposed of in Strabo’s narrative, whereas Semiramis’ city plays a central part in the first half of 16.1. Even though Ninus/Nineveh was “much greater” it is gone from Strabo’s history in a pair of quick sentences. 851 Strabo 16.1.5 (Babylon); 16.1.9-10 (Babylonian canals). 852 See below , p.247. 853 Strabo 16.1.5 (the wonders of Babylon); 16.1.15 (Babylon as metropolis of Assyria); 2.1.26 (the W all of Semiramis); 2.1.31 (foundation of Babylon and construction of the royal palace). 854 Strabo 16.1.5, see below , p.240. 855 Strabo 16.1.16. 245/448 Strabo refers to Ninus and Semiramis as masters of Asia (οὗτοι δὲ ἐκράτησαν τῆς Ἀσίας). 856 This role, and Semiramis’s expeditions and conquests in her own right, were a considerable part of the traditional Semiramis legend. Strabo mentions her expedition against India in his discussion of the difficulties of trusting prior authors on India. 857 However, he refers to it in order to dismiss it, citing Megasthenes’ opinion that no one had invaded India before Alexander. Strabo refers to the stories of Semiramis’ conquests in methodological sections, but never links them specifically to locations outside Assyria. Within Strabo’s narrative, Semiramis is confined to the border spaces where Parthian power was most visible to a Roman 856 Strabo 16.1.2. 857 Strabo 15.1.5-6: “Εἰ τοίνυν ταῦτ’ ἀφείς τις τὴν πρὸ τῆς Ἀλεξάνδρου στρατείας ἐπιβλέποι μνήμην, πολὺ ἂν εὕροι τούτων τυφλότερα. Ἀλέξανδρον μὲν οὖν πιστεύειν τοῖς τοιούτοις εἰκός, τετυφωμένον ταῖς τοσαύταις εὐτυχίαις. Φησὶ γοῦν Νέαρχος φιλονεικῆσαι αὐτὸν διὰ τῆς Γεδρωσίας ἀγαγεῖν τὴν στρατιάν, πεπυσμένον διότι καὶ Σεμίραμις ἐστράτευσεν ἐπὶ Ἰνδοὺς καὶ Κῦρος· ἀλλ’ ἡ μὲν ἀνέστρεψε φεύγουσα μετὰ εἴκοσιν ἀνθρώπων, ἐκεῖνος δὲ μεθ’ ἑπτά· ὡς σεμνὸν τὸ ἐκείνων τοσαῦτα παθόντων αὐτὸν καὶ στρατόπεδον διασῶσαι μετὰ νίκης διὰ τῶν αὐτῶν ἐθνῶν τε καὶ τόπων. [6] Ἐκεῖνος μὲν δὴ ἐπίστευσεν· ἡμῖν δὲ τίς ἂν δικαία γένοιτο πίστις περὶ τῶν Ἰνδικῶν ἐκ τῆς τοιαύτης στρατείας τοῦ Κύρου ἢ τῆς Σεμιράμιδος; συναποφαίνεται δέ πως καὶ Μεγασθένης τῷ λόγῳ τούτῳ κελεύων ἀπιστεῖν ταῖς ἀρχαίαις περὶ Ἰνδῶν ἱστορίαις· οὔτε γὰρ παρ’ Ἰνδῶν ἔξω σταλῆναί ποτε στρατιάν, οὔτ’ ἐπελθεῖν ἔξωθεν καὶ κρατῆσαι πλὴν τῆς μεθ’ Ἡρακλέους καὶ Διονύσου καὶ τῆς νῦν μετὰ Μακεδόνων. καίτοι Σέσωστριν μὲν τὸν Αἰγύπτιον καὶ Τεάρκωνα τὸν Αἰθίοπα ἕως Εὐρώπης προελθεῖν, Ναβοκοδρόσορον δὲ τὸν παρὰ Χαλδαίοις εὐδοκιμήσαντα Ἡρακλέους μᾶλλον καὶ ἕως Στηλῶν ἐλάσαι. μέχρι μὲν δὴ δεῦρο καὶ Τεάρκωνα ἀφικέσθαι, ἐκεῖνον δὲ καὶ ἐκ τῆς Ἰβηρίας εἰς τὴν Θρᾴκην καὶ τὸν Πόντον ἀγαγεῖν τὴν στρατιάν· Ἰδάνθυρσον δὲ τὸν Σκύθην ἐπιδραμεῖν τῆς Ἀσίας μέχρι Αἰγύπτου· τῆς δὲ Ἰνδικῆς μηδένα τούτων ἅψασθαι· καὶ Σεμίραμιν δ’ ἀποθανεῖν πρὸ τῆς ἐπιχειρήσεως· Πέρσας δὲ μισθοφόρους μὲν ἐκ τῆς Ἰνδικῆς μεταπέμψασθαι Ὑδράκας, ἐκεῖ δὲ μὴ στρατεῦσαι, ἀλλ’ ἐγγὺς ἐλθεῖν μόνον ἡνίκα Κῦρος ἤλαυνεν ἐπὶ Μασσαγέτας.” (If, however, one should dismiss these accounts and observe the records of the country prior to the expedition of Alexander, one would find things still more obscure. Now it is reasonable to suppose that Alexander believed such records because he was blinded by his numerous good fortunes; at any rate, Nearchus says that Alexander conceived an ambition to lead his army through Gedrosia when he learned that both Semiramis and Cyrus had made an expedition against the Indians, and that Semiramis had turned back in flight with only twenty people and Cyrus with seven; and that Alexander thought how grand it would be, when those had met with such reverses, if he himself should lead a whole victorious army safely through the same tribes and regions. Alexander, therefore, believed these accounts.[6] But as for us, what just credence can we place in the accounts of India derived from such an expedition made by Cyrus, or Semiramis? And Megasthenes virtually agrees with this reasoning when he bids us to have no faith in the ancient stories about the Indians; for, he says, neither was an army ever sent outside the country by the Indians nor did any outside army ever invade their country and master them, except that with Heracles and Dionysus and that in our times with the Macedonians. However, Sesostris, the Aegyptian, he adds, and Tearco the Aethiopian advanced as far as Europe; and Nabocodrosor, who enjoyed greater repute among the Chaldaeans than Heracles, led an army even as far as the Pillars. Thus far, he says, also Tearco went; and Sesostris also led his army from Iberia to Thrace and the Pontus; and Idanthyrsus the Scythian overran Asia as far as Aegypt; but no one of these touched India, and Semiramis too died before the attempt; and, although the Persians summoned the Hydraces as mercenary troops from India, the latter did not make an expedition to Persia, but only came near it when Cyrus was marching against the Massagetae.) Loeb trans. cf. Strabo repeats Alexander’s sentiment at 15.2.5. 246/448 viewer. Strabo deploys the same technique in his use of Alexander and the Macedonians in book 16. Alexander and the Macedonians Strabo’s treatment of Alexander and the Macedonian conquest also focuses the reader’s attention on the past. In this case, that past comprises a Hellenistic history of imperial expansion which became a central pillar of imperial ideologies in the geographical areas it touched until well into the Roman period and beyond. 858 Alexander’s expansion brought with it a great deal of physical mobility , not just in terms of the conquering armies, but of including the colonisation and administrative movement that followed. Macedonian colonisation is evident throughout Strabo’s description of Assyria. As we saw in Chapter 3, Strabo’s description of this region populated mostly by Aramaic speakers abounds with links to Greek culture. Ἔστι δ’ ἡ μὲν παρόρειος εὐδαίμων ἱκανῶς· ἔχουσι δ’ αὐτῆς τὰ μὲν πρὸς τῷ Εὐφράτῃ καὶ τῷ ζεύγματι, τῷ τε νῦν τῷ κατὰ τὴν Κομμαγηνὴν καὶ τῷ πάλαι τῷ κατὰ τὴν Θάψακον, οἱ Μυγδόνες κατονομασθέντες ὑπὸ τῶν Μακεδόνων· ἐν οἷς ἐστιν ἡ Νίσιβις, ἣν καὶ αὐτὴν Ἀντιόχειαν τὴν ἐν τῇ Μυγδονίᾳ προσηγόρευσαν, ὑπὸ τῷ Μασίῳ ὄρει κειμένην, καὶ Τιγρανόκερτα καὶ τὰ περὶ Κάρρας καὶ Νικηφόριον χωρία... 859 Strabo knows this region as Mygdonia, a name given by the Macedonians. As previously discussed, Mygdonis had resonances of in mythic Greek history and the Macedonian landscape. 860 Many of the sites within that space had multiple links to Greek and Macedonian culture. Zeugma, a Greek work meaning “link”, “bridge” or “crossing”, was named for the important crossing of the Euphrates. It was founded as a Macedonian settlement on the west bank of the river by Seleucus and bore the alternate name Seleucia. Seleucus also founded the city on the opposite bank named Apamea after his wife. Mygdonian Antiocheia, 858 “The conquests of Alexander the Great and the work of his successors had long ago planted the landscape of Mesopotamia with Greek cities, which were within the intellectual horizons of men of the Classical world, even when they were outside its political boundaries.” Matthews (1989) 140. 859 Strabo 16.1.23: “The country alongside the mountains is quite fertile; the parts of it near the Euphrates and Zeugma... are occupied by the Mygdones, who were so named by the Macedonians. In their country lies Nisibis, which is also called Mygdonian Antiocheia; it lies at the foot of Mt. Masius, and so do Tigranocerta and the lands of Carrhae and Nicephorium...” 860 See Chapter 1. 247/448 the Macedonian name for the important city of Nisibis, was one of many Antiochs founded by Seleucid kings – Antiochus was a popular dynastic name. Nikephorium is also a Greek name, “bearer of victory”, although the founder is uncertain; Seleucus I is most likely , although some sources credit Alexander himself, or Seleucus II. 861 While Strabo does not go into detail about these Macedonian settlements, their very presence in the landscape and their Greek and Macedonian names in Strabo’s narrative gives recognizable evidence of cultural contact between former settlers and local people. 862 By infusing the landscape he describes with Macedonian elements, Strabo establishes and prioritises the links between the space of northern Mesopotamia and its Hellenistic history . This prioritisation of Hellenistic elements fits with Strabo’s traditionally Greek understanding of the antithetical relationship between the cultural categories “Greek” and “Barbarian”. 863 This scheme was complicated by the strength and ubiquity of Roman power; a problem with which his predecessor Polybius had also grappled. 864 Strabo resolved the potential difficulty by placing Rome in the position of civiliser and bearer of Greek knowledge to the barbarians. 865 As culturally superior, the Greeks occupied a privileged position within the world Strabo constructed. In Mesopotamia this emerged in the foundational role of Macedonian colonisation and in the links he established between the area and the Greek mythic past. 866 Moreover, these Macedonian places have contemporary Roman resonances for Strabo’s readers. 861 Pliny (NH. 6.119) and Isidore (1) credit Alexander, Appian (Syr. 57) credits Seleucus I, while the Syriac chronicle of the twelfth-century Jacobite patriarch Michael the Syrian (4.78, Michael and Chabot (1963); S. P . Brock (1979) 15–17) and the anonymous Greek Chronicon Paschale (J.P .Migne (ed) Patrologiae Cursus Completeus, Series Graeca 92 (o.J.) Col 429f.) of the seventh century both credit Seleucus II with the foundation. Kessler, Karlheinz, ‘Nicephorium’ BNP 9.714; al-Khalaf and Kohlmeyer (1985). 862 Strabo’s narrative gives no information about the degree of segregation or integration of these groups, but regardless, some interaction must have occurred. 863 Dueck (2000) 75; Almagor (2005) 43. 864 Walbank (1972) 160–66; Momigliano (1975) 29–31, 37–39; Dueck (2000) 52–53. Strabo regarded himself as a continuator of Polybius and the relationship between the two authors is much studied. Two recent works with much to say in this regard are Clarke (1999) and Dueck (2000). In particular, note the presence of Polybius in Strabo’s lists of worthy predecessors (1.1.1; 1.2.1). 865 Dueck (2000) 79–84; Almagor (2005) 53. For the civilising influence of Roman conquest, see Strabo 3.2.15; 4.1.12. 866 Strabo (16.1.25) links the Gordyaeans to Gordys, the son of Triptolemus. 248/448 Zeugma remained the most important crossing of the Euphrates well into the Roman period and was periodically the site of a legionary base, as was Samosata in Commagene. 867 Carrhae was the site of one of the most famous defeats in Roman military history in 53 BCE, Nikephorion was Crassus’s most important base in that campaign against the Parthians, and Sinnaca was the site of that general’s death. Nisibis may have been too far east to enter into Roman affairs much by Strabo’s time. Strabo’s historical notes usually place notable episodes in local history into a Roman context. 868 This is easily accomplished for the majority of his work as much of the area covered had been conquered and incorporated into the Roman empire. Beyond the eastern extent of Roman territory , Strabo frequently used Alexander of Macedon, a figure with a powerful hold on the Roman imagination, to provide familiar context and to establish a link between the reader and the subject matter. 869 Early in book 16, Strabo referred to Alexander’s victory at Gaugamela: “In Aturia is a village Gaugamela, where Dareius was conquered and lost his empire” (ἐν δὲ τῇ Ἀτουρίᾳ ἐστὶ Γαυγάμηλα κώμη, ἐν ᾗ συνέβη νικηθῆναι καὶ ἀποβαλεῖν τὴν ἀρχὴν Δαρεῖον). 870 Although Alexander himself is not mentioned, the name Gaugamela would have been well- known to Strabo’s educated readers. In ascribing the loss of Darius’ empire to the outcome of this battle, Strabo drew a historical boundary between Persian and Macedonian histories and Iranian and Hellenistic Greek culture. To connect Alexander and Assyria to Rome, Strabo drew upon the Roman celebration of their own national talent for governance and organisation. 871 867 Edwell (2008) 18; Kennedy (1998) 139–62; Appendix 1.20. 868 Although his account of Greece and western Anatolia is more closely linked to Homer. Syme described these portions as recycled Homeric lecture notes. 869 Many of Strabo’s sources dated to the time of Alexander, or wrote in contexts in which there was political or cultural capital in linking the deeds of Alexander to people and places (This practice of Alexander’s contemporaries was censured by Eratosthenes: Arr. Anab. 5.3.1; BNJ 241 F28 = Plut. Alex. 3.2; cf. Strabo 15.1.7- 8). However, while Alexander’s conquests had opened the east to Hellenistic geography , it was not mandatory to mention the Macedonian conqueror in relation to every site which he visited. When Strabo mentions Alexander, we should assume that he did so intentionally. 870 Strabo 16.1.3. 871 Vir. Aen. 6.851-53: “tu regere imperio populos, Romane, memento / (hae tibi erunt artes), pacique imponere morem, / parcere subiectis et debellare superbos.” (You, Roman, remember to rule the people with authority , (for these are your arts), to impose the custom of peace, to spare the subjected and vanquish the arrogant.) 249/448 Most of Strabo’s references to Alexander concern his brief administrative record in Babylonia. He describes how the Macedonian demolished the artificial cataracts on the Euphrates and Tigris rivers which the Persians had constructed, according to Strabo, out of fear of attack. 872 He also notes that Alexander took care to inspect and maintain the elaborate canal system of Babylonia, with the implication that the Persians had not done so. 873 Strabo presents these works as opening communication between Babylonia and the surrounding world, including Arabia. 874 The latter was particularly important for Alexander as he was apparently planning a military expedition against the Arabians before his death. 875 These activities of a 872 Strabo 16.1.9: “Πέρσαι τοὺς ἀνάπλους ἐπίτηδες κωλύειν θέλοντες φόβῳ τῶν ἔξωθεν ἐφόδων καταράκτας χειροποιήτους κατεσκευάκεισαν· ὁ δὲ Ἀλέξανδρος ἐπιὼν ὅσους οἷός τε ἦν ἀνεσκεύασε, καὶ μάλιστα τοὺς ἐπὶ τὴν Ὦπιν.” (Now the Persians, wishing on purpose to prevent voyaging up these rivers, for fear of attacks from without, had constructed artificial cataracts, but Alexander, when he went against them, destroyed as many of them as he could, and in particular those to Opis.) Loeb trans. 873 Strabo 16.1.9: “ἐπεμελήθη δὲ καὶ τῶν διωρύγων· πλημμυρεῖ γὰρ ὁ Εὐφράτης κατὰ τὴν ἀρχὴν τοῦ θέρους ἀπὸ τοῦ ἔαρος ἀρξάμενος, ἡνίκα τήκονται αἱ χιόνες αἱ ἀπὸ τῆς Ἀρμενίας, ὥστ’ ἀνάγκη λιμνάζειν καὶ κατακλύζεσθαι τὰς ἀρούρας, εἰ μὴ διοχετεύει τις ταφρείαις καὶ διώρυξι τὸ ἐκπῖπτον τοῦ ῥοῦ καὶ ἐπιπολάζον ὕδωρ, καθάπερ καὶ ἐν Αἰγύπτῳ τὸ τοῦ Νείλου.” (He also paid careful attention to the canals; for the Euphrates rises to flood-tide at the beginning of summer, beginning first to rise in the spring when the snows in Armenia melt; so that of necessity it forms lakes and deluges the ploughed lands, unless the excess of the stream, or the surface water, is distributed by means of trenches and canals, as is the case with the Nile in Egypt.) Loeb trans. Strabo 16.1.11: “Φησὶ δ’ Ἀριστόβουλος τὸν Ἀλέξανδρον αὐτὸν ἀναπλέοντα καὶ κυβερνῶντα τὸ σκάφος ἐπισκοπεῖν καὶ ἀνακαθαίρειν τὰς διώρυγας μετὰ τοῦ πλήθους τῶν συνακολουθησάντων· ὡς δ’ αὕτως καὶ τὰ στόμια ἐμφράττειν, τὰ δ’ ἀνοίγειν· κατανοήσαντα δὲ μίαν τὴν μάλιστα τείνουσαν ἐπὶ τὰ ἕλη καὶ τὰς λίμνας τὰς πρὸ τῆς Ἀραβίας, δυσμεταχείριστον ἔχουσαν τὸ στόμα καὶ μὴ ῥᾳδίως ἐμφράττεσθαι δυναμένην διὰ τὸ εὐένδοτον καὶ μαλακόγειον, ἄλλο ἀνοῖξαι καινὸν στόμα, ἀπὸ σταδίων τριάκοντα ὑπόπετρον λαβόντα χωρίον, κἀκεῖ μεταγαγεῖν τὸ ῥεῖθρον· ταῦτα δὲ ποιεῖν προνοοῦντα ἅμα καὶ τοῦ μὴ τὴν Ἀραβίαν δυσείσβολον τελέως ὑπὸ τῶν λιμνῶν ἢ καὶ τῶν ἑλῶν ἀποτελεσθῆναι, νησίζουσαν ἤδη διὰ τὸ πλῆθος τοῦ ὕδατος... ταῦτά τε δὴ πραγματεύεσθαι περὶ τὰς διώρυγας τὸν Ἀλέξανδρον, καὶ τοὺς τάφους σκευωρεῖσθαι τοὺς τῶν βασιλέων καὶ δυναστῶν· τοὺς γὰρ πλείστους ἐν ταῖς λίμναις εἶναι.” (Aristobulus says that Alexander himself, when he was sailing up the river and piloting the boat, inspected canals and with his multitude of followers cleared them; and that he likewise stopped up some of the mouths and opened others; and when he noticed that one canal, the one which stretched most directly towards the marshes and lakes that lay in front of Arabia, had a mouth most difficult to deal with and could not easily be stopped up because of the yielding and soft nature of the soil, he opened up another mouth, a new one, at a distance of thirty stadia from it, having selected a place with a rocky bottom, and that he diverted the stream to that place; and that in doing this he was taking forethought at the same time that Arabia should not be made utterly difficult to enter by the lakes or even by the marshes, since, on account of the abundance of water, that country was already taking the form of an island... Accordingly , he adds, Alexander busied himself thus with the canals, and also inspected thoroughly the tombs of the kings and potentates, most of which are situated among the lakes.) Loeb trans. Another example of Alexander repairing structures neglected by the Persians, 16.1.5. 874 Although the “cataracts” were probably important parts of the irrigation system of Babylonia, Bosworth (1988) 159. 875 Bosworth (1988) 168–70. Strabo 16.1.11: “διανοεῖσθαι γὰρ δὴ κατακτᾶσθαι τὴν χώραν ταύτην καὶ στόλους 250/448 successful ruler would have had particular resonance to a Roman audience familiar with the ideological stance of Augustus, Rome’s first emperor, as the re-builder of Rome and re-organiser of the empire following the neglect of the late republic. Augustus and the ideological vision of Rome and Roman power were important themes in Strabo’s work. 876 Strabo refers and alludes to Augustus frequently throughout his Geography , creating a picture of the Roman emperor as a conqueror, patron, builder, peacemaker, and philhellene. 877 Just as Strabo never enters into a full biography of Augustus, so he never articulates a full comparison between Augustus and Alexander. Nevertheless, on several occasions their achievements are compared in passing or juxtaposed. 878 In his description of Egypt in book 17, Strabo writes at some length on the negligence of the Ptolemiac dynasty and the reorganisation of that state after its conquest by Rome. Strabo concludes his description of the misrule of the last Ptolemies and Roman republican military dynasts with a note that after the defeat of Antony and Cleopatra at the battle of Actium “Augustus Caesar pursued and destroyed both, and stopped the drunken violence in Egypt.” 879 This is immediately followed by a note on the new status of Egypt καὶ ὁρμητήρια ἤδη κατεσκευάσθαι, τὰ πλοῖα τὰ μὲν ἐν Φοινίκῃ τε καὶ Κύπρῳ ναυπηγησάμενον διάλυτά τε καὶ γομφωτά, ἃ κομισθέντα εἰς Θάψακον σταθμοῖς ἑπτὰ εἶτα τῷ ποταμῷ κατακομισθῆναι μέχρι Βαβυλῶνος, τὰ δ’ ἐν τῇ Βαβυλωνίᾳ συμπηξάμενον τῶν ἐν τοῖς ἄλσεσι καὶ τοῖς παραδείσοις κυπαρίττων· σπάνις γὰρ ὕλης ἐνταῦθα, ἐν δὲ Κοσσαίοις καὶ ἄλλοις τισὶ μετρία τίς ἐστιν εὐπορία. σκήψασθαι μὲν οὖν αἰτίαν τοῦ πολέμου φησίν, ἐπειδὴ μόνοι τῶν ἁπάντων οὐ πρεσβεύσαιντο οἱ Ἄραβες ὡς αὐτόν, τὸ δ’ ἀληθὲς ὀρεγόμενον πάντων εἶναι κύριον· καὶ ἐπεὶ δύο θεοὺς ἐπυνθάνετο τιμᾶσθαι μόνους ὑπ’ αὐτῶν, τόν τε Δία καὶ τὸν Διόνυσον, τοὺς τὰ κυριώτατα πρὸς τὸ ζῆν παρέχοντας, τρίτον ὑπολαβεῖν ἑαυτὸν τιμήσεσθαι, κρατήσαντα καὶ ἐπιτρέψαντα τὴν πάτριον αὐτονομίαν ἔχειν ἣν εἶχον πρότερον.” (For of course Alexander, he says, intended to acquire possession of that country , and had already prepared fleets and bases of operations, having built some of his boats in Phoenicia and Cypros, boats that were constructed with bolts and could be taken to pieces, which were conveyed by a seven days' journey to Thapsacus and then down the river to Babylon, and having built others in Babylonia, from the cypress trees in the groves and the parks; for there is a scarcity of timber in Babylonia, although there is a moderately good supply of timber in the countries of the Cossaei and certain other tribes. Now Alexander alleged as cause of the war, Aristobulus says, that the Arabians were the only people on earth who did not send ambassadors to him, but in truth was reaching out to be lord of all; and when he learned that they worshipped two gods only , Zeus and Dionysus, the gods who supply the most requisite needs of life, he took it for granted that they would worship him as a third if he mastered them and allowed them to keep the ancestral independence which they had had before.) Loeb trans. 876 Dueck (2000) 85–106; Nicolet (1991) 20–24. 877 Dueck (2000) 96–106 collates Strabo’s references to Augustus and his deeds. 878 Dueck (2000) 104; especially Strabo 1.2.1. 879 Strabo 17.1.11: “καὶ μετὰ ταῦτα ἐπακολουθήσας ὁ Σεβαστὸς Καῖσαρ ἀμφοτέρους κατέλυσε καὶ τὴν 251/448 (“ἐπαρχία δὲ νῦν ἐστι”), then a long passage on the orderly and prudent rule of officials appointed by Augustus. 880 The idea that Roman power and conquest had opened up communications with insular regions had an important place in Strabo’s work. The finale of his account of the Roman reorganisation of Egypt is a description of the increased trade and prosperity that had resulted from that reorganisation. Just as Strabo describes Alexander reorganising the water routes of Babylonia and removing the artificial impediments of a former empire from the rivers, so he has Roman power under Augustus reorganising Egypt’s political and administrative boundaries, removing former impediments and opening its wealth to the world. 881 Strabo explicitly focuses on trade beyond the bounds of the Roman Empire in his description of the Egyptian case, writing that large fleets now ply the routes through the Red Sea to India and Arabia. 882 Representing the Parthians While Strabo is interested in historical matters, he is primarily concerned with the present and practical implications of his work. 883 However, in his description of “Assyria” in book 16, Strabo reduced the present rulers of the region, namely the Arsacid dynasty of Parthia, to a relatively minor role. They appear in three places: in Strabo’s description of Ctesiphon, in his description of the various oppressors of northern Mesopotamia, and in a discussion of the Romano-Parthian border at the end of his description of Assyria. In each of these locations, the Parthian role in the space under discussion is vague and underplayed. In book 16, Strabo first mentions the Parthians in his treatment of the cities at the heart of Parthian- controlled Babylonia. After describing Babylon extensively , he describes its decline at the expense of Seleucia Αἴγυπτον ἔπαυσε παροινουμένην.” 880 Strabo 17.1.12: “It is now a province”. Strabo 17.1.12-13 describes the Roman organisation of Egypt and its virtues, Polybius' distaste for the Ptolemaic organisation and the Roman encouragement of trade. 881 Strabo 17.1.5. 882 Chapter five examines the emphasis on trade in descriptions of northern Mesopotamia. 883 Strabo 1.1.23; 2.5.13; Dueck (2000) 51–52; Nicolet (1991) 73. For Strabo’s use of Polybius as a model in this regard, see Dueck (2000) 47–48; especially regarding Polyb. 3.7.4-7. 252/448 on the Tigris, the important Macedonian foundation and administrative centre. 884 Strabo introduces Seleucia as a great replacement to Babylon early in book 16; however this is a misleading description of Seleucia’s status in Strabo’s day . It is only much later that he describes the contemporary situation: Πάλαι μὲν οὖν ἡ Βαβυλὼν ἦν μητρόπολις τῆς Ἀσσυρίας, νῦν δὲ Σελεύκεια ἡ ἐπὶ τῷ Τίγρει λεγομένη. πλησίον δ’ ἐστὶ κώμη Κτησιφῶν λεγομένη, μεγάλη· ταύτην δ’ ἐποιοῦντο χειμάδιον οἱ τῶν Παρθυαίων βασιλεῖς φειδόμενοι τῶν Σελευκέων, ἵνα μὴ κατασταθμεύοιντο ὑπὸ τοῦ Σκυθικοῦ φύλου καὶ στρατιωτικοῦ· δυνάμει οὖν Παρθικῇ πόλις ἀντὶ κώμης ἐστὶ καὶ τὸ μέγεθος, τοσοῦτόν γε πλῆθος δεχομένη καὶ τὴν κατασκευὴν ὑπ’ ἐκείνων αὐτῶν κατεσκευασμένη καὶ τὰ ὤνια καὶ τὰς τέχνας προσφόρους ἐκείνοις πεπορισμένη. εἰώθασι γὰρ ἐνταῦθα τοῦ χειμῶνος διάγειν οἱ βασιλεῖς διὰ τὸ εὐάερον· θέρους δὲ ἐν Ἐκβατάνοις καὶ τῇ Ὑρκανίᾳ διὰ τὴν ἐπικράτειαν τῆς παλαιᾶς δόξης. 885 Strabo’s narrative is remarkable. He begins with a contrast between the two cities he had previously mentioned (Babylon and Seleucia) signalled grammatically by a μὲν … δὲ construction and chronologically by contrasting temporal adverbs (Πάλαι … νῦν). Immediately after this, he adds the Parthian capital Ctesiphon as a third site, located nearby (πλησίον). Then he gradually gives further information about the Parthian capital of Mesopotamia: it is a village (κώμη), a large one (μεγάλη), the winter quarters (χειμάδιον) of the Parthian kings, it is a city rather than a village (πόλις ἀντὶ κώμης). He goes on to describe how the 884 Strabo 16.1.5: ἀλλὰ καὶ τὰ λοιπὰ ὠλιγωρήθη καὶ κατήρειψαν τῆς πόλεως τὰ μὲν οἱ Πέρσαι τὰ δ’ ὁ χρόνος καὶ ἡ τῶν Μακεδόνων ὀλιγωρία περὶ τὰ τοιαῦτα, καὶ μάλιστα ἐπειδὴ τὴν Σελεύκειαν ἐπὶ τῷ Τίγρει πλησίον τῆς Βαβυλῶνος ἐν τριακοσίοις που σταδίοις ἐτείχισε Σέλευκος ὁ Νικάτωρ. καὶ γὰρ ἐκεῖνος καὶ οἱ μετ’ αὐτὸν ἅπαντες περὶ ταύτην ἐσπούδασαν τὴν πόλιν καὶ τὸ βασίλειον ἐνταῦθα μετήνεγκαν·καὶ δὴ καὶ νῦν ἡ μὲν γέγονε Βαβυλῶνος μείζων ἡ δ’ ἔρημος ἡ πολλή, ὥστ’ ἐπ’ αὐτῆς μὴ ἂν ὀκνῆσαί τινα εἰπεῖν ὅπερ ἔφη τις τῶν κωμικῶν ἐπὶ τῶν Μεγαλοπολιτῶν τῶν ἐν Ἀρκαδίᾳ “ἐρημία μεγάλη ’στὶν ἡ Μεγάλη πόλις.” (...what was left of [Babylon] was neglected and thrown into ruins, partly by the Persians and partly by time and by the indifference of the Macedonians to things of this kind, and in particular after Seleucus Nicator had fortified Seleuceia on the Tigris near Babylon, at a distance of about three hundred stadia therefrom. For not only he, but also all his successors, were strongly interested in Seleuceia and transferred the royal residence to it. What is more, Seleuceia at the present time has become larger than Babylon, whereas the greater part of Babylon is so deserted that one would not hesitate to say what one of the comic poets said in reference to the Megalopolitans in Arcadia: ‘The Great City is a great desert.’) Loeb trans. Strabo also uses this comic line of Megalopolis, 8.8.1. The poet is unknown. 885 Strabo 16.1.16: “Long ago Babylon was the metropolis of Assyria; but now Seleuceia on the Tigris has that name. Nearby is a village called Ctesiphon, a large one. The kings of the Parthians made this their winter residence, sparing the Seleuceians, so that the Seleuceians might not be oppressed by the requirements of billeting Scythian people and soldiers. Because of Parthian power, Ctesiphon is a city rather than a village; its size is such that it receives a multitude of people, and it has been equipped with buildings and furnished with goods and arts suitable to them. For the kings are accustomed to spend the winter there because of the good air, and the summer at Ecbatana and in Hyrcania because of the superiority of their ancient glory.” 253/448 Parthian presence has resulted in a large and important city . Strabo delays this begrudging description of the Parthian capital until halfway through the chapter and then he reveals that description piecemeal as if unwilling to acknowledge it. He then immediately follows the description by mentioning Babylonia and Seleuceia, the two great cities which he had already introduced near the beginning of the chapter. His arrangement of the chapter on Assyria emphasises the importance of Babylon and Seleucia and only later, and then grudgingly and briefly , acknowledges Ctesiphon as an important political centre. This construction serves to make light of Parthian power in Babylonia. It is typical of Strabo’s treatment of the Parthians in this section. The second case where the Parthians are mentioned comes amidst one of Strabo’s descriptions of the Arabian Skenitai, tent-dwelling nomads of the north Syrian desert. Here Strabo notes that: τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα τῆς Μεσοποταμίας καὶ ἀπωτέρω τῶν ὀρῶν ἄνυδρα καὶ λυπρὰ ὄντα ἔχουσιν οἱ σκηνῖται Ἄραβες, λῃστρικοί τινες καὶ ποιμενικοί, μεθιστάμενοι ῥᾳδίως εἰς ἄλλους τόπους, ὅταν ἐπιλείπωσιν αἱ νομαὶ καὶ αἱ λεηλασίαι. τοῖς οὖν παρορείοις ὑπό τε τούτων κακοῦσθαι συμβαίνει καὶ ὑπὸ τῶν Ἀρμενίων· ὑπέρκεινται δὲ καὶ καταδυναστεύουσι διὰ τὴν ἰσχύν· τέλος δ’ ὑπ’ ἐκείνοις εἰσὶ τὸ πλέον ἢ τοῖς Παρθυαίοις· ἐν πλευραῖς γάρ εἰσι κἀκεῖνοι τήν τε Μηδίαν ἔχοντες καὶ τὴν Βαβυλωνίαν. 886 The Skenitae are the focus of the section. Strabo emphasises the fluidity of their mobility , a way of life with pejorative implications in the sedentary worldview of Greek and Roman culture. 887 They are bandits (λῃστρικοί) and shepherds (ποιμενικοί), characteristics which Strabo emphasises by explicating the conditions of their mobility: a lack of pasturage (αἱ νομαὶ) or victims to rob (αἱ λεηλασίαι). He then describes their victims, the people near the mountains (τοῖς παρορείοις), that is, the Mygdonians. The 886 Strabo 16.1.26: “The parts of Mesopotamia which incline towards the south and are farther from the mountains, which are waterless and barren, are occupied by the Arabian Skenitai, a tribe of brigands and shepherds, who readily move from one place to another when pasture and booty fail them. Accordingly , the people who live alongside the mountains are harassed not only by the Scenitae, but also by the Armenians, who are situated above them and, through their might, oppress them; and at last they are subject for the most part to the Armenians or else to the Parthians, for the Parthians too are situated on the sides of the country and possess both Media and Babylonia.” 887 Shaw (1995). 254/448 unfortunate Mygdonians are also harassed by other neighbouring groups: the Armenians and the Parthians. This is the second occasion on which Parthian power is mentioned in Strabo’s Assyria. Here the Parthian empire is placed in an equal position to the Armenians as oppressors (καταδυναστεύουσι) of the Mygdonian cities (with all their Hellenistic markers). The placement of this exertion of Parthian power in a section on nomadic banditry further minimises its force as a stable source of authority in the region. Parthian power in northern Mesopotamia is implicitly compared to, and placed on the same level as, the power of Armenia, a Roman client state, or that of the Skenitai, characterised as little more than bandits. Strabo depicts contemporary Assyria as an ambiguous borderland space where political boundaries are elided and obscured. His description of Mesopotamia ends with a discussion of the border between the Romans and the Parthians. Ὅριον δ’ ἐστὶ τῆς Παρθυαίων ἀρχῆς ὁ Εὐφράτης καὶ ἡ περαία· τὰ δ’ ἐντὸς ἔχουσι Ῥωμαῖοι καὶ τῶν Ἀράβων οἱ φύλαρχοι μέχρι Βαβυλωνίας, οἱ μὲν μᾶλλον ἐκείνοις οἱ δὲ τοῖς Ῥωμαίοις προσέχοντες, οἷσπερ καὶ πλησιόχωροί εἰσιν, ἧττον μὲν Σκηνῖται οἱ νομάδες οἱ τῷ ποταμῷ πλησίον, μᾶλλον δ’ οἱ ἄπωθεν καὶ πρὸς τῇ εὐδαίμονι Ἀραβίᾳ. 888 Strabo writes that the Euphrates is the “boundary of the Parthian empire” (Ὅριον δ’ ἐστὶ τῆς Παρθυαίων ἀρχῆς). He makes no overt statement about any territorial limit to Roman power. He goes on to say that the Syrian side of the Euphrates, even as far as Babylonia, is held by the Romans and nomadic Arab chieftains, although some of those chieftains are more inclined towards the Parthians and others toward the Romans generally based on their proximity to each. Thus Strabo explicitly defines a limit for the Parthians (which he then implicitly shows them exceeding) while he does not place a limit on Roman power, instead leaving the area which Rome controlled mostly implicit. This leads the reader to suppose that the reality of Roman power extended further than it did in practice. Moreover, by not stating a boundary to Roman power, it 888 Strabo 16.1.28: “The Euphrates and the land beyond it constitute the boundary of the Parthian Empire. But the parts this side of the river are held by the Romans and the chieftains of the Arabians as far as Babylonia, some of the chieftains preferring to give ear to the Parthians and others to the Romans, to whom they are neighbours; less so the nomad Skenitai who are near the river, but more so those that are far away and near Arabia Felix.” 255/448 implies that it has no boundary , thus reinforcing the Augustan ideology of universal dominion. 889 By showing the Parthians implicitly exceeding their limit (ὅριον), Strabo also creates an ambiguous border wherein both empires could exercise power. Following this description of the border between the Parthians and Romans, the rest of the chapter is devoted to a historical account of recent border relations between the two empires: οἱ δὲ Παρθυαῖοι καὶ πρότερον μὲν ἐφρόντιζον τῆς πρὸς Ῥωμαίους φιλίας, τὸν δὲ ἄρξαντα πολέμου Κράσσον ἠμύναντο· καὶ αὐτοὶ ἄρξαντες τῆς μάχης τῶν ἴσων ἔτυχον, ἡνίκα ἔπεμψαν ἐπὶ τὴν Ἀσίαν Πάκορον ... (Ἀντώνιος δὲ συμβούλῳ τῷ Ἀρμενίῳ χρώμενος προὐδόθη καὶ κακῶς ἐπολέμησεν)· ὁ δ’ ἐκεῖνον διαδεξάμενος Φραάτης τοσοῦτον ἐσπούδασε περὶ τὴν φιλίαν τὴν πρὸς Καίσαρα τὸν Σεβαστὸν ὥστε καὶ τὰ τρόπαια ἔπεμψεν ἃ κατὰ Ῥωμαίων ἀνέστησαν Παρθυαῖοι, καὶ καλέσας εἰς σύλλογον Τίτιον τὸν ἐπιστατοῦντα τότε τῆς Συρίας, τέτταρας παῖδας γνησίους ἐνεχείρισεν ὅμηρα αὐτῷ, Σερασπαδάνην καὶ Ῥωδάσπην καὶ Φραάτην καὶ Βονώνην, καὶ γυναῖκας τούτων δύο καὶ υἱεῖς τέτταρας, δεδιὼς τὰς στάσεις καὶ τοὺς ἐπιτιθεμένους αὐτῷ· ᾔδει γὰρ μηδένα ἰσχύσοντα καθ’ ἑαυτόν, ἂν μή τινα ἐπιλάβῃ τοῦ Ἀρσακίου γένους διὰ τὸ εἶναι σφόδρα φιλαρσάκας τοὺς Παρθυαίους· ἐκποδὼν οὖν ἐποίησε τοὺς παῖδας, ἀφελέσθαι ζητῶν τὴν ἐλπίδα ταύτην τοὺς κακουργοῦντας. τῶν μὲν οὖν παίδων ὅσοι περίεισιν ἐν Ῥώμῃ δημοσίᾳ βασιλικῶς τημελοῦνται· καὶ οἱ λοιποὶ δὲ βασιλεῖς πρεσβευόμενοι καὶ εἰς συλλόγους ἀφικνούμενοι διατετελέκασιν. 890 In this brief and selective account of foreign affairs in the region, Strabo places the blame for the conflict 889 For Strabo’s view that the limits of Roman power correspond to the oikoumene, see Dueck (2000) 109–11 For the relationship between Strabo and Augustan poetry , see Dueck (2000) 123–4 On Strabo as a spiritually Augustan work, see Dueck (2000) 159–60. In his discussion of the rise of Rome at 6.4.2, Strabo states that Rome and Parthia share a common border, but he does not give that impression in the discussion of the borderland itself. 890 Strabo 16.1.28: “In earlier times, the Parthians gave heed to Roman friendship. They defended themselves against Crassus, who began war with them; and then, having begun the battle themselves, suffered the same fate when they sent Pakorus against Asia. (Antony , using the Armenian as an adviser, was betrayed and suffered badly in his war.) Phraates, his successor, was so eager for the friendship of Caesar Augustus that he even sent him the trophies which the Parthians had set up to celebrate their victory over the Romans. And, having called Titius (then governor of Syria) to a conference, he put into his hands four of his legitimate sons as hostages, Seraspadanes and Rhodaspes and Phraates and Bonones, and two wives and four sons of these, fearing civil strife and attacks against him. For he knew that no one could prevail against him unless that person supported a descendant of Arsaces, because the Parthians were extremely loyal to the Arsacids. Thus, he put his children out of the way , seeking to take away that hope from evil-doers. Now as many of his children as remain are cared for in Rome, in royal style and at public expense. and the kings which followed him have continued to send ambassadors and attend meetings." There are two textual problems with this passage. 1) Letronne and Groskurd both include a clause following “ἡνίκα ἔπεμψαν ἐπὶ τὴν Ἀσίαν Πάκορον” which briefly explicates the heritage of Pakorus, Radt (2002) 4.306. 2) I follow Radt in punctuating the following line with brackets: “Ἀντώνιος δὲ συμβούλῳ τῷ Ἀρμενίῳ χρώμενος προὐδόθη καὶ κακῶς ἐπολέμησεν”. This simply clarifies that Phraates’ succession (διαδεξάμενος Φραάτης) proceeds from the Parthian kings, rather than from Antonius or the Armenians, the actors in the intervening clause. 256/448 between the empires on Crassus, a suitable scapegoat as both a rival of Augustus’ divine father Julius Caesar and as a Roman general whose spectacular military failure against the Parthians was repaired by Augustus’ diplomatic efforts. Moreover, the historical tradition had attached several religious errors to Crassus’ campaign, allowing an implicit contrast to be drawn with Augustus’s revival of traditional Roman religious propriety . 891 To compound Crassus’ failure and emphasise Augustus’ success further, Strabo inserts a parenthetical condemnation of Antonius, Augustus’ personal rival (Ἀντώνιος... κακῶς ἐπολέμησεν). As for the Parthians their own attempts at aggressive warfare against Roman territory also failed (τῶν ἴσων ἔτυχον). This presentation of the period gives the Parthians agency , but emphasises the mutual inability of the two empires to expand. Furthermore, Strabo twice states the desire of Parthian kings to establish friendship (φιλίας) with Rome: “οἱ Παρθυαῖοι... ἐφρόντιζον τῆς πρὸς Ῥωμαίους φιλίας” and “ὁ Φραάτης τοσοῦτον ἐσπούδασε περὶ τὴν φιλίαν...”. This use of φιλία may reflect the Roman language of amicitia and clientship. 892 By making the Parthians the active party in seeking friendship in both cases, Strabo places them in the position of supplicant, subordinates them politically and minimises Parthian power in relation to Rome. This was in keeping with Augustan propaganda around the return of the legionary standards lost by Crassus at Carrhae. The return of Crassus’ lost legionary standards was a cornerstone of the Augustan ideology of eastern domination. Strabo’s treatment of the affair is entirely in keeping with this perspective. The Parthian king Phraates IV (r. 37-2 BCE) was so eager for Roman friendship that he handed over four legitimate sons, two wives and their unnumbered sons as hostages (ὅμηρα) who are now all cared for at Rome (τῶν μὲν οὖν 891 Beard et al. (1998). 892 Badian (1997); Burton (2011) summarises the relevant scholarship since Badian. Augustus’ account of the affair in the Res Gestae (29) uses the word amicita, which is translated in the Greek text as φιλία. Res Gestae 29: “Parthos trium exercitum Romanorum spolia et signa reddere mihi supplicesque amicitiam populí Romaní petere coegi.” “Πάρθους τριῶν στρατευμάτῶν Ῥωμαίων σκῦλα καὶ σημέας ἀποδοῦναι ἐμοὶ ἱκέτας τε φιλίαν δήμου Ῥωμαίων ἀξιῶσαι ἡνάγκασα.” (The Parthians I compelled to restore to me the spoils and standards of three Roman armies, and to seek as suppliants the friendship of the Roman people.) 257/448 παίδων ὅσοι περίεισιν ἐν Ῥώμῃ δημοσίᾳ βασιλικῶς τημελοῦνται). 893 The bare facts of the affair were certainly open to the Augustan representation which subordinated the Parthian kings to Roman authority , but Strabo’s notice of the hostage giving reveals the likely truth which lay beneath. When he gave over these hostages to Rome, Phraates IV had ruled for over three decades and was looking to his succession. In sending these sons and wives into Roman power, he cleared the way for his young successor, the future Phraates V . 894 Potential contenders to the Arsacid throne always came from within the Arsacid house: Phraates relatives were either potential rivals in their own right or potential pawns of rival non-Arsacid factions among other Parthian noble families. Phraates himself had killed his father and thirty brothers to claim the throne and ensure his succession. Having ensured the succession of his favoured son, Phraates now succumbed to the same fate. Giving them into Roman power not only limited the role they could play in Arsacid palace politics, it isolated them from Parthian affairs and tainted them with foreign upbringing and potential loyalty , reducing their threat to Phraates and his chosen succession. This treatment of Parthian power and control in Mesopotamia was not based in ignorance on Strabo’s part. He certainly knew about the Parthians and displayed some of that knowledge in his discussion of Parthia itself. 895 In that narrative, he gives a few paragraphs of description then excuses himself from writing more about the Parthians: “But since I have said much about Parthian usages in the sixth book of my Historical Sketches and in the second book of my History of events after Polybius, I shall omit discussion of that subject here, lest I may seem to be repeating what I have already said.” (εἰρηκότες δὲ πολλὰ περὶ τῶν Παρθικῶν νομίμων ἐν τῇ ἕκτῃ τῶν ἱστορικῶν ὑπομνημάτων βίβλῳ, δευτέρᾳ δὲ τῶν μετὰ Πολύβιον, παραλείψομεν ἐνταῦθα μὴ ταυτολογεῖν δόξωμεν). 896 Unfortunately Strabo’s historical works are lost, so our 893 Bivar (2000) 67–68. 894 Bivar (2000) 68. 895 Strabo 11.9.1-3. 896 Strabo 11.9.3: “εἰρηκότες δὲ πολλὰ περὶ τῶν Παρθικῶν νομίμων ἐν τῇ ἕκτῃ τῶν ἱστορικῶν ὑπομνημάτων βίβλῳ, δευτέρᾳ δὲ τῶν μετὰ Πολύβιον, παραλείψομεν ἐνταῦθα μὴ ταυτολογεῖν δόξωμεν, τοσοῦτον εἰπόντες μόνον ὅτι τῶν Παρθυαίων συνέδριόν φησιν εἶναι Ποσειδώνιος διττόν, τὸ μὲν συγγενῶν τὸ δὲ σοφῶν καὶ 258/448 examination of Strabo’s knowledge of the Parthians is limited, but the fragmentary picture of them that we receive from the Geography suggests that he knew more about Parthian history and administration in Assyria than he provides in that chapter. At the beginning of his work he credits the spread of both the Roman and Parthian empires for an increased availability of geographical knowledge. 897 In the Hellenistic and Roman context, much of this geographic knowledge was accompanied by historical descriptions of political and military activity at the edges of the empire. The same was probably true for the various lost Parthian histories. Strabo writes that more recent writing about peoples and places on the fringes of the known world provides better information than that of the past. In particular, he notes “the writers of Parthian histories, Apollodorus of Artemita and his school” (“ὑπὸ τῶν τὰ Παρθικὰ συγγραψάντων, τῶν περὶ Ἀπολλόδωρον τὸν Ἀρτεμιτηνόν”). 898 Little is known of Apollodorus of Artemita. He probably wrote a μάγων, ἐξ ὧν ἀμφοῖν τοὺς βασιλεῖς καθίστασθαι.” (But since I have said much about Parthian usages in the sixth book of my Historical Sketches and in the second book of my History of events after Polybius, I shall omit discussion of that subject here, lest I may seem to be repeating what I have already said, though I shall mention this alone, that the Council of the Parthians, according to Poseidonius, consists of two groups, one that of kinsmen, and the other that of wise men and Magi, from both of which groups the kings were appointed.) Loeb trans. On Strabo’s other works, see Dueck (2000) 69–72. As well as his comments on the Parthians, something of the scope of his historical works is shown by fragments preserved in Josephus. 897 Strabo 1.2.1: “καὶ γὰρ δὴ πολύ τι τοῖς νῦν ἡ τῶν Ῥωμαίων ἐπικράτεια καὶ τῶν Παρθυαίων τῆς τοιαύτης ἐμπειρίας προσδέδωκε, καθάπερ τοῖς μετὰ τὴν Ἀλεξάνδρου στρατείαν, ὥς φησιν Ἐρατοσθένης· ὁ μὲν γὰρ τῆς Ἀσίας πολλὴν ἀνεκάλυψεν ἡμῖν καὶ τῶν βορείων τῆς Εὐρώπης ἅπαντα μέχρι τοῦ Ἴστρου· οἱ δὲ Ῥωμαῖοι τὰ ἑσπέρια τῆς Εὐρώπης ἅπαντα μέχρι Ἄλβιος ποταμοῦ τοῦ τὴν Γερμανίαν δίχα διαιροῦντος τά τε πέραν Ἴστρου τὰ μέχρι Τύρα ποταμοῦ· τὰ δὲ ἐπέ κεινα μέχρι Μαιωτῶν καὶ τῆς εἰς Κόλχους τελευτώσης παραλίας Μιθριδάτης ὁ κληθεὶς Εὐπάτωρ ἐποίησε γνώριμα καὶ οἱ ἐκείνου στρατηγοί· οἱ δὲ Παρθυαῖοι τὰ περὶ τὴν Ὑρκανίαν καὶ τὴν Βακτριανὴν καὶ τοὺς ὑπὲρ τούτων Σκύθας γνωριμωτέρους ἡμῖν ἐποίησαν, ἧττον γνωριζομένους ὑπὸ τῶν πρότερον· ὥστε ἔχοιμεν ἄν τι λέγειν πλέον τῶν πρὸ ἡμῶν.” (Indeed, the spread of the empires of the Romans and of the Parthians has presented to geographers of today a considerable addition to our empirical knowledge of geography , just as did the campaign of Alexander to geographers of earlier times, as Eratosthenes points out. For Alexander opened up for us geographers a great part of Asia and all the northern part of Europe as far as the Ister River; the Romans have made known all the western part of Europe as far as the River Albis (which divides Germany into two parts), and that regions beyond the Ister as far as the Tyras River; and Mithridates, surnamed Eupator, and his generals have made known the regions beyond the Tyras as far as Lake Maeotis and the line of coast that ends at Colchis; and, again, the Parthians have increased our knowledge in regard to Hyrcania and Bactriana, and in regard to the Scythians who live north of Hyrcania and Bactriana, all of which countries were but imperfectly known to the earlier geographers. I therefore may have something more to say than my predecessors.) Loeb trans. 898 Strabo 2.5.12: “Μάλιστα δ’ οἱ νῦν ἄμεινον ἔχοιεν ἄν τι λέγειν περὶ τῶν κατὰ Βρεττανοὺς καὶ Γερμανοὺς καὶ τοὺς περὶ τὸν Ἴστρον τούς τε ἐντὸς καὶ τοὺς ἐκτὸς Γέτας τε καὶ Τυρεγέτας καὶ Βαστάρνας, ἔτι δὲ τοὺς περὶ τὸν Καύκασον, οἷον Ἀλβανοὺς καὶ Ἴβηρας. ἀπήγγελται δ’ ἡμῖν καὶ ὑπὸ τῶν τὰ Παρθικὰ συγγραψάντων, τῶν περὶ Ἀπολλόδωρον τὸν Ἀρτεμιτηνόν, ἃ πολλῶν ἐκεῖνοι μᾶλλον ἀφώρισαν, τὰ περὶ τὴν Ὑρκανίαν καὶ 259/448 Parthika in at least four books during the 1 st century BCE which described the rise of the Arsacids and the expansion of Parthian power from the Iranian plateau. 899 Ten of the twelve extant fragments of his work recorded in Brill’s New Jacoby are preserved in Strabo and all of them deal with places and events east of the Zagros mountains, except for a fragment concerning the Iberi and Armenia. 900 This is not just an issue of the sources available to Strabo, but an issue of how he selected material to highlight or obscure details in response to contemporary political currents. Sarah Pothecary has shown how Strabo obscured Germanicus’ role in settling affairs in Armenia, Cappadocia and Commagene (all of which Strabo describes) out of respect for Tiberius’ wish that Germanicus not be glorified after his death. 901 Strabo had a large quantity of material at his disposal from which he selected carefully to construct the space to fit his understanding of the borderland and the events within it. Strabo’s Geography relies heavily on a legacy of Hellenistic knowledge and action, but the world he presents is centred on Rome. In his Geography , Strabo orients his geographical description of Assyria towards τὴν Βακτριανήν· τῶν τε Ῥωμαίων καὶ εἰς τὴν εὐδαίμονα Ἀραβίαν ἐμβαλόντων μετὰ στρατιᾶς νεωστί, ἧς ἡγεῖτο ἀνὴρ φίλος ἡμῖν καὶ ἑταῖρος Αἴλιος Γάλλος, καὶ τῶν ἐκ τῆς Ἀλεξανδρείας ἐμπόρων στόλοις ἤδη πλεόντων διὰ τοῦ Νείλου καὶ τοῦ Ἀραβίου κόλπου μέχρι τῆς Ἰνδικῆς, πολὺ μᾶλλον καὶ ταῦτα ἔγνωσται τοῖς νῦν ἢ τοῖς πρὸ ἡμῶν. ὅτε γοῦν Γάλλος ἐπῆρχε τῆς Αἰγύπτου, συνόντες αὐτῷ καὶ συναναβάντες μέχρι Συήνης καὶ τῶν Αἰθιοπικῶν ὅρων ἱστοροῦμεν, ὅτι καὶ ἑκατὸν καὶ εἴκοσι νῆες πλέοιεν ἐκ Μυὸς ὅρμου πρὸς τὴν Ἰνδικήν, πρότερον ἐπὶ τῶν Πτολεμαϊκῶν βασιλέων ὀλίγων παντάπασι θαρρούντων πλεῖν καὶ τὸν Ἰνδικὸν ἐμπορεύεσθαι φόρτον.” (In particular the writers of the present time can give a better account of the Britons, the Germans, the peoples both north and south of the Ister, the Getans, the Tyregetans, the Bastarnians, and, furthermore, the peoples in the regions of the Caucasus, such as the Albanians and the Iberians. Information has been given us also concerning Hyrcania and Bactriana by the writers of Parthian histories (Apollodorus of Artemita and his school), in which they marked off those countries more definitely than many other writers. Again, since the Romans have recently invaded Arabia Felix with an army , of which Aelius Gallus, my friend and companion, was the commander, and since the merchants of Alexandria are already sailing with fleets by way of the Nile and of the Arabian Gulf as far as India, these regions also have become far better known to us of today than to our predecessors. At any rate, when Gallus was prefect of Egypt, I accompanied him and ascended the Nile as far as Syene and the frontiers of Ethiopia, and I learned that as many as one hundred and twenty vessels were sailing from Myos Hormos to India, whereas formerly , under the Ptolemies, only a very few ventured to undertake the voyage and to carry on traffic in Indian merchandise.) Loeb trans. 899 Gärtner, Hans Armin, “Apollodorus [8a]” BNP . D’Hautcourt, Alexis, “Apollodoros of Artemita (779)”, BNJ; Alonso-Núñez (1989); Nikonorov (1998). 900 D’Hautcourt, Alexis, “Apollodoros of Artemita (779)”, BNJ. See D’Hautcourt’s commentary for the problems with identifying Apollodorus’ material. The Armenian fragment is BNJ 779 F 2 = Strabo 1.3.21. 901 Pothecary (2002). 260/448 past political power rather than current political power, adapts the Semiramis legend to this ideological purpose, and configures Semiramis as a foundational figure in the region, literally and figuratively . Alexander and the Macedonians play a similar role by drawing attention to a past which had been co-opted into Roman conceptions of imperial glory and geographical expansion. Strabo’s narrative prioritises Hellenism, and even pre-Hellenistic cultures, at the expense of the Parthians. When he does treat Parthian power in the region directly , he minimises it and subordinates it to that of Rome, implicitly showing Rome as the dominant force in a border region where their military ambitions had been repeatedly thwarted. Pliny (Dis)ordering Space Pliny’s description of the area of the Mesopotamian borderland is split between two separate books. As we saw in Chapter 1, his description of “Syria” includes a section reflecting an itinerary down the Euphrates and a description of Osrhoene and other areas east of that river, while his description of the area between the Euphrates and Tigris rivers comes as part of his description of Asia in book 6. 902 These two narratives overlap in northwest Mesopotamia: both mention the Orroei, Anthemus/Anthemusia and Nikephorion. 903 While Pliny never explicitly defines the border in his own time, his differential treatment of the two spaces—Syria and Mesopotamia—reveals his conception of the two spheres of imperial power and implies a political division. 904 Through his selection of material, delimitation of space and boundaries, and the construction of his narrative, Pliny portrays Provincia Syria (broadly conceived) as an ordered Roman space and the bulk of Mesopotamia as a chaotic Parthian space. Book Five: Syria Pliny’s account of the space of ancient Syria is quite geographically coherent. In the Natural History 902 Euphrates itinerary: NH 5.86-87. Osrhoene et al.: NH 5.85-86. Mesopotamia: NH 6.117-120. 903 Orroei: NH 5.85, 6.117. Anthemus(ia) and Nikephorion: NH 5.86; 6.118. 904 He defines the border in the time of Pompey (NH 6.120) and notes that Philiscum is a Parthian city as if to indicate part of a border (NH 5.89). 261/448 networks of water provide structure for description of bodies of land. 905 The geographic books are organised according to the periplus genre of coastal voyages, and the core of his geography may have come from one such source. 906 Throughout his work, the narrative constantly and explicitly returns to the coast or to rivers. 907 Accordingly , in his description of Syria, his narrative moves up the coast, then down the Euphrates, with occasional diversions into the interior regions as he proceeds. He begins this narrative trail at Pelusium where he left his account of Egypt and proceeds up the coast with diversions to describe the inland areas of Judaea, Phoenecia and Syria. 908 After reaching the north end of the Levantine coast he lists a number of interior cities between Antioch and the Euphrates. 909 This is followed by a description of the course of the Euphrates and the adjacent regions of Commagane and Osrhoene. 910 Where the Euphrates bends east at Sura, Pliny’s narrative leaves the river to describe Palmyra and other desert cities in southern Syria before returning to the Euphrates to complete the river’s course to the salt-marshes at the head of the Persian Gulf. 911 This largely linear approach gives structure to Pliny’s geographical narrative, but does not limit it. As mentioned above, Pliny diverges from his periplus style to describe areas not obviously tied to a route, but these descriptions are somewhat list-like and geographically disconnected and remain so until Pliny’s narrative can return to the safety of a shoreline. At times the narrative makes no attempt at relative spatial arrangement and instead orders space according to other criteria; for example, in his inland description 905 Murphy (2004) 135–6. 906 Nicolet (1991) 173–4. However, note W oolf on Pliny’s deliberate choice to shape a periplus from source narratives in a variety of styles, W oolf (2011) 11. That Pliny structures his geographical narrative on a periplus around the Mediterranean quickly becomes clear as one reads his work. He begins with a description of the coast of Europe as comprising four gulfs (Pliny NH 3.1.). At a closer scale this is particularly clear in his account of the Peloponnese in which lists of sites are grouped according to the bays to which they are closest (see especially 4.16-19 on the sinus Laconicus, sinus Argolicus, and sinus Saronicus). 907 A few examples: Pliny , NH 4.81-88; 5.66, 68, 75; 6.77-8. 908 Pliny NH 68-79. 909 Pliny NH 81-2. 910 Pliny , NH, 5.88-90. Quoted below , p.239. 911 Pliny , NH, 5.86-7. 262/448 linking the coast with the Euphrates, Pliny lists cities of the interior without topological links: Nunc interiora dicantur. Coele habet Apameam, Marsya amne divisam a Nazerinorum tetrarchia, Bambycen, quae alio nomine Hierapolis vocatur, Syris vero Mabog — ibi prodigiosa Atargatis, Graecis autem Derceto dicta, colitur —, Chalcidem cognominatam Ad Belum, unde regio Chalcidena fertilissima Syriae, et inde Cyrresticae Cyrrum, Gazetas, Gindarenos, Gabenos, tetrarchias duas quae Granucomatitae vocantur, Hemesenos, Hylatas, Ituraeorum gentem et qui ex his Baethaemi vocantur, Mariamnitanos, [82] tetrarchiam quae Mammisea appellatur, Paradisum, Pagras, Penelenitas, Seleucias praeter iam dictam duas, quae Ad Euphraten et quae Ad Belum vocantur, Tardytenses. reliqua autem Syria habet, exceptis quae cum Euphrate dicentur, Arethusios, Beroeenses, Epiphanenses ad Orontem, Laodicenos, qui Ad Libanum cognominantur, Leucadios, Larisaeos, praeter tetrarchias in regna discriptas barbaris nominibus XVII. 912 Although the narrative disguises the fact with parenthetical notes about some of the locations mentioned, it is clear that Pliny is working from two alphabetical lists of locations and tribes, one for Coele Syria (“Coele habet... Tardytenses”) and another for the rest of Syria (“reliquam autem Syria habet... barbaris nominibus XVII”). However, these two lists do not cover distinct geographical areas (as Map 23: Pliny's Syria shows). 913 Rather, the second list (reliqua Syria) is a list of ethnic descriptors, while the first comprises mostly cities. Jones proposed that Pliny used an official list of Syrian peoples which he combined with the places mentioned by his literary sources as belonging in Coele Syria to compose his first alphabetised list; his second list comprised the remaining peoples from that official list. 914 The inclusion of narrative passages that are essentially lists of places without topological relations between them is explicitly in keeping with Pliny’s project, as outlined in the introduction to his geographical section: Locorum nuda nomina et quanta dabitur brevitate ponentur, claritate causisque dilatis in suas partes; 912 Pliny , NH, 5.81-82: “Let the interior now be described. Coele Syria has Apamea, separated from the tetrarchy of the Nazerini by the river Maysa; Bambyce (which his also called Hierapolis, and in fact Mabog by the Syrians – here the monstrous Atargatis is worshipped, called Derceto by the Greeks) Chalcis, named On the Belus, from which the very fertile region of Syria, Chalcidene, is named, and then Cyrrus of Cyrrestica, the Gazetae, Gindareni, Gabeni and the two tetrarchies called Granucomatitae, the Hemeseni, Hylatae, the tribe of Ituraei (and one branch called the Baethaemi), the Mariamnitani, [82] the tetrarchy called Mammisea, Paradisum, Pagrae, Penelenitae, two Seleucias (beyond that already mentioned), one called “on the Euphrates” the other “on the Belus”, and the Tardytenses. The rest of Syria, except for the parts to be addressed with the Euphrates, has the Arethusii, the Beroeenses, the Epiphanenses on the Orontes, the Laodiceans (called “on Lebanon”), the Leucadii, the Larisaeos, and seventeen more tetrarchies with barbarian names divided into kingdoms.” 913 The locations of the sites and peoples on the list are discussed in Appendix 1.2. 914 Jones (1971) 260–63. 263/448 nunc enim sermo de toto est. 915 As Murphy has argued, the organisational principle of Pliny’s work is to categorise the world opened up to and organised by Roman power. 916 In his geographical books, Pliny is primarily concerned with listing the “bare names of places” (locorum nuda nomina). Sometimes it suits his purposes to place these names in topological relation with each other, at others it does not. Because Pliny seeks to give an overview de toto, the specific geographical relationships are often unimportant to his project. However, that they be named and listed as subordinate parts of the Roman empire is valuable to his conception of the whole. 915 Pliny NH 3.1-2: “The bare names of places will be set down, and with the greatest brevity available, their celebrity and its reasons being deferred to the proper sections; for my topic now is the world as a whole.” 916 Murphy (2004). 264/448 Map 23: Pliny's Syria Pliny’s attention to Roman power and its display is evident in his treatment of the passage of the Euphrates through the Taurus range. 917 As noted in Chapter 1, this passage interrupts the geographical flow of the narrative and dislocates the reader from the area of Samosata to southern Cappadocia, before the narrative continues down the river. 918 The evocative martial language of the brief digression (pugna, victus, vincit, fractum, expellit) is heralded and perhaps inspired by Pliny’s use of expugnat for the emergence of the Euphrates from the mountains. This kind of digression is common in Pliny . Murphy has argued that Pliny’s loose sentence structure and digressive narrative are an aesthetic choice. 919 Pliny’s text contains frequent asides mentioning his concern to discuss famous and interesting locations and has a sense of playfulness and intricacy that livens what could be a rather dull progression of facts. This kind of digression acted as a demonstration of Pliny’s literary skills. Murphy discusses the martial language which surrounds and frames this passage in the context of the eternal battle between rivers and mountains that plays out throughout Pliny’s work. 920 This timeless, amphitheatrical spectacle, in which two anonymised combatants play out their respective roles for the pleasure of the dominating gaze of the elite Roman audience, falls within the realm of a literary aesthetic focused on placing the world in relation to its Roman masters. Book Six: Mesopotamia W e have seen how Pliny’s descriptions of the Mesopotamian borderland are constructed as lists. In book 5, these are a coastal periplus (5.75-80), two atopological (and alphabetical) lists in northern Syria (5.81-82), and the Euphrates itinerary (5.83-87, 89-90), the last interrupted by a list of desert ports (5.88-89). 917 Pliny NH 5.85: “apud Claudiopolim Cappadociae cursum ad occasum solis agit. primo hunc illic in pugna Taurus aufert victusque et abscisus sibimet alio modo vincit ac fractum expellit in meridiem. ita naturae dimicatio illa aequatur, hoc eunte quo vult, illo prohibente ire qua velit.” (At Claudiopolis in Cappadocia [the Euphrates] turns its course to the west. There, the Taurus redirects the river for the first time in this conflict and despite being conquered and cut in half it wins in another way and expels it, broken, to the south. Thus this struggle of nature is a draw , the river breaking through as it wished, but the mountain stopping its desired course). 918 See the map and discussion of Pliny’s Syria in Chapter 1. 919 Murphy (2004) 29–40. 920 Murphy (2004) 148–151. French also discusses theatricality in Pliny , (1994) 202–3. 265/448 These lists follow a clear route up the Syrian coast, across northern Syria and down the Euphrates. Where the location of a list item is unclear, it is often possible to locate it in relation to neighbouring list items. The exception to this are the two alphabetical lists of people and places in northern Syria. However, while the lists in book 5 are arranged in a topologically coherent manner for the most part, those in book 6 are jumbled together without apparent structure. In book 6, the lists are the “Arab” tribes of Mesopotamia (6.117-18), the mountain tribes between Gordyene and the highlands of Adiabene (6.118), a list of unlocated towns (6.118), and the itinerary from Zeugma to Pompey’s Roman border (6.119-20). Essentially , Pliny gives a list of subordinate lists. Within three of those lists, sites and people can be located relative to each other. In the fourth, the list of unlocated towns are linked by common Macedonian denomination rather than topological references, as is Nikephorion which follows. For this reason, several of the places and people that Pliny mentions are difficult to identify , especially when named only by a common dynastic toponym. “Seleucia” could refer to Zeugma (Seleucia on the Euphrates) or Seleucia on the Tigris. Pliny probably means the former as the latter is described in some detail in the context of Babylonia. 921 However, Pliny’s does not maintain a strict division between his use of the terms Babylonia and Mesopotamia and his narrative jumps from place to place on several occasions in his description of the region. 922 The only Artemita known in the region is the city east of the Tigris, described by Isidore as πολιν ‘Ελληνιδα ʼΑρτεμιτα. 923 No towns called Laodicea are known in Mesopotamia. Because of this topological incoherence, the four lists are not able to be related to each other spatially based only on the internal evidence of the text; knowledge external to the text (such as personal knowledge or a map) is required. Although Roman and Parthian space is never clearly delimited in the Natural History (the only place in the borderland clearly defined as Parthian is the oppidum of Philiscum on the Euphrates), 924 the division 921 Pliny NH 6.122. 922 At NH 6.123 Pliny describes several Babylonian oppida as “in Mesopotamia”. 923 Isidore 2; Schoff (1989) 6. See above, p.239. 924 Pliny , NH 5.89: “a Sura autem proxime est Philiscum, oppidum Parthorum ad Euphraten.” See Appendix 3. 266/448 of the geographical narrative between two books strongly implies that the reader is to take that part described in book five as Roman space and that part described in book six as Parthian. In book five, Syria appears after Egypt and before Cilicia in a circumlocution of the Roman Mediterranean; in book six, Mesopotamia appears after the Iranian plateau (including an explicit, and brief, description of the Parthian empire) and before Babylonian and Arabia. The different narrative structures of the two sections give the “Roman” part a greater sense of coherence and organisation than the “Parthian”. However, the lack of explicit notice of territorial divisions makes the presence of the division between Roman and Parthian space ambiguous. The linear internal links in these lists, combined with the overall lack of links between these lists gives a general impression of disorder. However, it is uncertain how many of Pliny’s readers would have possessed sufficient geographic knowledge of Mesopotamia to recognise these discontinuities. This disorder might only have been perceived by someone familiar with the space under discussion. Nevertheless, perceptive readers would still detect the alphabetical categorisation of Roman power evident in the north Syrian lists and the structure of nested lists in the Mesopotamian narrative, even if the spatial dislocation did not emphasise the disorder. This embedded linear representation of the world conforms to Pliny’s organisational principle of categorising and displaying the world under Roman power. 925 In The Edges of the Earth in Ancient Thought, James Romm argues that catalogues of exotic wonders generate their own authority through a matter-of-fact tone, simple assertion of existence, and density of information. 926 The 925 Murphy (2004); Naas (2011); French (1994) 208–10. 926 “...the matter-of-fact tone of the catalogue helps balance the exoticism of the wonders themselves: The insistence on spare linguistic structures such as the simple assertion of existence, “there is,” creates a veneer of clinical dispassionate inquiry. Furthermore, the frequent citation of sources which typifies this kind of writing also helps bolster authority—or at least relieves the author himself, and thereby his audience, of responsibility for assessing truth. Beyond these strategies, however, the very density of the catalogue plays a crucial role in its effectiveness, replicating at a stylistic level the density of the Indian landscape itself. The cataloguer of wonders typically crowds his data into linguistic aggregates rather than discrete units: prolix, hyperextended sentences, or rapid-fire series of clauses. The proliferation of data dislocates its readers and overwhelms their ability to separate true from false...” Romm (1992) 91ff. For the relationship between imperialism and wonder catalogues in Pliny specifically , see Naas (2011). 267/448 rhetorical technique that Romm describes applies to the Natural History broadly , but also applies to the kind of geographical narrative that Pliny presents here. That geographical narrative is full of long sentences, which, although grammatically simple, threaten to overwhelm the reader with lists of brief clauses and bare names. In Pliny’s description, the list is less the medium of presentation than it is the message itself. Spatial relationships between provincial settlements are seldom important to Pliny . What is important is the relationship of dominance between the imperial centre and the many settlements which parade through the Natural History. A reader unfamiliar with the spatial realities of the Mesopotamian borderland could be impressed with the coherence of a list of Roman subject cities and peoples in north Syria, and still recognise the incoherent and shifting nature of the series of lists of peoples and places not subject to Roman organisation. Imperial Space Geography and imperialism are intertwined. The ability to conceptualise and organise space is required for conquest and government. 927 This can be seen at the local level in the Roman practice of centuriation, discussed in chapter three. 928 The collection, aggregation and presentation of geographic information at the start of the Roman empire came in a time of reorganisation at the end of a period of widespread internal conflict. The imperial geographical writers handle the division of imperial space and the representation of political boundaries differently . In Strabo’s time, Roman power had not penetrated Mesopotamia very far. Accordingly , Strabo’s usual method of presenting a world dominated by Roman political and economic power was inadequate for this region. He could have modified his procedure of linking Roman space to Roman power by simply linking Parthian space to Parthian power, but instead he looked to the past for alternative methods of representing a 927 Nicolet (1991) 2. Ando (2000) 277: “Imperialism possesses its own logic and requires a particular geography.” 928 For the relationship between centuriation and the colonisation of the landscape, see Purcell (1990). 268/448 space unbound to Roman power. He looked to the Greek historical and literary past for cultural connections with resonance to a Greek audience or to a Roman audience familiar with the Greek tradition. The cultural world of a Hellenised urban elite were familiar territory to Strabo, and to his readers, and in the absence of Roman political power, he linked the spaces he described to the “trans-national” Hellenic cultural power which Rome had adopted as her own. Literary links to a Greek tradition of adapting Semiramis, historical links to Alexander of Macedon, cultural links to the Macedonian Seleucids, and a political narrative that obscured Parthian power at the expense of Roman ideology provided a substitute structure for Strabo’s presentation of the Mesopotamian borderland. Strabo represents the internal political and administrative divisions of Roman space as well as the provincial organisation of the world overall. He provides the clearest indication of the separation of Roman and Parthian space by including an explicit notification that the Euphrates formed such a border. 929 Strabo’s narrative also supports a kind of intellectual boundary on the Euphrates. His descriptions of the internal divisions of the broader regions he describes become steadily less detailed and specific as his narrative deals with places further beyond the river. Commagene and other internal Syrian sub-divisions are reasonably clear with regional names associated with cities therein, Mesopotamia becomes more fuzzy , and Adiabene is confused. This increasingly confused representation of extra-Roman space is found to an even greater extent and in a more stark division in Pliny’s work. Pliny reflects the new era of political control in Provincia Syria by portraying Syria as an ordered Roman space and the geographic area of Mesopotamia as a chaotic and ambiguously Parthian space. 930 Mesopotamia is discussed in two geographically overlapping sections and for the most part comprises multiple lists poorly anchored to geographic features and presented in a spatially disjointed narrative that blends into Assyria and Babylonia without concern for regional boundaries. By 929 Strabo 16.1.28. 930 This dichotomy between Roman order and non-Roman chaos is also seen in Strabo’s Geography , but not in his description of Mesopotamia. Dueck (2000) 65–66. 269/448 contrast, Pliny’s representation of Syria consists of a spatially linked itinerary around a defined area, albeit with the inland regions described in a pair of spatially confused lists. Nevertheless, the “Syria” described in book 5 is a coherent unit in a way that the space east of the Euphrates is not. Within that former unit, Pliny’s use of lists gives a sense of regimented organisation that his description of the latter does not. This contrast between the organised Roman world of Syria and the disorganised non-Roman world beyond the Euphrates reinforces the overlapping zone east of the Euphrates as a political boundary region. The less detailed accounts of subsequent geographical writers do not allow examination to the same extent as Strabo and Pliny , but their delineation of boundaries and imperial space can be seen to some degree. Ptolemy’s work is the least political, but it nevertheless represents internal political boundaries within Roman space. Although he does not make the distinction explicit in the text, his work is structured according to Roman provincial organisation within Roman territory , while the structure is geographic and ethnographic outside Roman territory . 931 Within Roman spaces like Syria, space is further divided along political or administrative lines. However, in extra-Roman spaces, the internal divisions by which the narrative is structured are geographical not political (for example, “near the Euphrates”, “near the Tigris” and so on). The Expositio describes the Roman world in a provincial structure and some of the non-Roman world as a list of adjacent tribes and regions with varying levels of description. 932 While the Expositio does not discuss a border between the two empires, it draws a sharp conceptual contrast between urbanised Roman lands (nostra terra) and the ethnically defined areas to the east. The narrative of the Expositio begins to the east and progresses towards Roman territory . The construction of this narrative suggests that the author expects that the border constitutes something of an information barrier to his audience. In fact, it seems that 931 Berggren and Jones (2000) 40–41. Although the MS tradition contains inconsistencies in the division of the provinces, Diller (1939). 932 The non-Roman areas: Expositio 8-20. 270/448 the author of the Expositio has a limited knowledge of or concern for areas beyond the bounds of the Roman Empire. He reports narrowly on economic factors to the east that may be useful to his audience. As the narrative approaches Roman space, territories are described in either geographic or cultural terms. Where a reader might not be expected to know the people or land in question, regio or gens is used with the proper name for clarity , but where knowledge of certain lands or peoples is assumed, the proper name stands alone. 933 As expected, the latter cases increase in frequency as the narrative draws closer to Roman space. For the author of the Expositio, space beyond the Roman Empire is imagined as a linear path through a series of geographic or cultural spaces, as a passing foreign traveller or trader might experience it. Through greater depth of information and relative complexity of topological information, the descriptions of the imperial provincia convey an impression of Roman order and prosperity not evident to the east. The source of this information is unclear, but it is not first hand. Just after his description of those extra-Roman territories, the author notes “and in fact, an historian (historicus) says this about the preceding peoples”. 934 This marks a break between information explicitly gathered from an unnamed source (the historicus), and information over which the Expositio implicitly claims a more expert knowledge. 935 When the narrative reaches the border, the Expositio pauses to give an authorial comment re-asserting his authority over the narrative and marking the ideological boundary between Persian lands (Persarum terras) and Roman territory (nostram terram). 936 Jean Rougé has cast doubt on the opening clause of the sentence containing the 933 Regio or gens: Expositio 9, 10, 11, 13, 14, 15, 17. Proper name alone: Expositio 16, 18, 19, 20, 22. 934 Expositio 21: “et haec quidem de praedictis gentibus historicus ait”. 935 In a minor way , the appearance of the historicus in the narrative marks a temporal shift from a historical to a contemporary perspective. Information about non-Roman space was related to the narrator by a historical researcher, which implies that it is information about a past state, compared to the narrator’s present position. However, this implication is not reinforced by the overall narrative or the grammar of the two parts (Roman and non-Roman). 936 Expositio 21: “Quoniam vero necessarium est et nostram terram, hoc est Romanorum, conscribere, experiar exponere, ut possit legentibus prodesse. Incipiamus ergo. † Diximus enim Persarum terras... †” (Since truly it is necessary to survey our land, that is the land of the Romans, I will attempt to explain so that it can be useful to readers. So we will begin. For we have described the lands of the Persians...) 271/448 plural terras (“diximus enim Persarum terras”). 937 If that plural form is genuine, then the juxtaposition of the unitary singular Roman terra against the divided and multiple Persian terrae fits the desired impression of the Expositio. In contrasting Roman order with Persian disorder, the Expositio reveals a significant continuity in Roman geographic thought over the first four centuries of the Roman Empire. In Chapter 3, we saw a strain of continuity in the conception of the civilising influence of Macedonian colonisation between Pliny and Ammianus Marcellinus. 938 Both authors introduce the borderland as a relatively backwards area which was improved through Macedonian urbanisation. Their conception of the borderland’s history was similar, but their representation of its contemporary status shows an important development over the intervening three centuries of increased Roman activity in the area. Whereas Pliny’s account of Mesopotamia presents a land implicitly divided between Hellenised cities and un-Hellenised tribal and pastoralist ethnic groups, Ammianus’ survey of Syria presents the most important cities of the region, regardless of their apparent Hellenism, and even mentions the pre-Hellenic foundation of several communities. 939 Moreover, Ammianus does not contrast the relative levels of organisation of the two 937 Rougé (1966) 154. 938 Pliny NH 6.117: “Mesopotamia tota Assyriorum fuit, vicatim dispersa praeter Babylona et Ninum. Macedones eam in urbes congregavere propter ubertatem soli”(All Mesopotamia belonged to the Assyrians, the population scattered in villages except for Babyon and Ninevah. The Macedonians gathered them together in cities because of the fertility of the soil.); Amm. Marc. 14.8.5-6: “orientis vero limes in longum protentus et rectum ab Euphratis fluminis ripis ad usque supercilia porrigitur Nili, laeva Saracenis conterminans gentibus, dextra pelagi fragoribus patens, quam plagam Nicator Seleucus occupatam auxit magnum in modum, cum post Alexandri Macedonis obitum successorio iure teneret regna Persidis, efficaciae inpetrabilis rex, ut indicat cognomentum. [6] Abusus enim multitudine hominum, quam tranquillis in rebus diutius rexit, ex agrestibus habitaculis urbes construxit multis opibus firmas et viribus, quarum ad praesens pleraeque licet Graecis nominibus appellentur, quae isdem ad arbitrium inposita sunt conditoris, primigenia tamen nomina non amittunt, quae eis Assyria lingua institutores veteres indiderunt.” (But the frontier of the East, extending a long distance in a straight line, reaches from the banks of the Euphrates to the borders of the Nile, bounded on the left by the Saracen peoples and open on the right to the waves of the sea. Seleucus Nicator took possession of the region and greatly increased it in power, when, after the death of Alexander of Macedon, he was holding the rule of Persia by right of succession; a capable and successful king, as his surname [Nicator] indicates. [6] For he took advantage of the great number of men whom he ruled for along time in peace, and built cities that were secure in wealth and power out of their rustic dwellings. Now most of these are called by the Greek names which their founder imposed upon them, nevertheless have not lost the original names which their ancient settlers gave them in the Assyrian language.) 939 For example, 14.8.9: “Post hanc acclinis Libano monti Phoenice, regio plena gratiarum et venustatis, urbibus decorata magnis et pulchris; in quibus amoenitate celebritateque nominum Tyros excellit, Sidon et Berytus 272/448 imperial domains. Ammianus gives a coherent picture of the political border between the Roman and Persian empires. He explicitly notes the point at which Julian’s army enters Sasanid space, emphasises the crossing with narrative elements and historiographical tropes, and supports that explicit notification with his description of the military activities of the army on either side of the border. 940 In 363 CE, the Mesopotamian border ran up the Khabur to the forts north of Hatra as far as Bezabde; explicit near the Euphrates where a geographical feature (the Khabur) allowed precision, less precise in the lands south and east of Nisibis. The internal divisions of the two empires are marked, but are not geographically specified as precisely as in other writers; for example, the boundaries between the provinces of Syria, Osrhoene and Mesopotamia are not defined. As far as we can tell from the extant geographical descriptions in his work, Ammianus also used the Euphrates as a regional boundary between Syria and Mesopotamia, but the historian was more concerned with political activities and boundaries than geographical description and boundaries. Although the form of his lost description of Mesopotamia is unknown, in the extant portion of his work Ammianus never explicitly describes the geographical borders of Mesopotamia and refers more often to Roman administrative categories than geographical regions. Moreover, in Sasanid territory , he considers all of Assyria as one region, bounded by the Niphates and Zagros mountains to the north and east and by Roman territory to the west. Conclusion Geographic writing can tend to be treated as an objective description of space, but the processes of selection, generalisation and aggregation that underlie geographic projects are inherently subjective. 941 While eisdemque pares Emissa et Damascus saeculis condita priscis.” (After this comes Phoenicia, lying at the foot of Mount Libanus, a region full of charm and beauty , adorned with many great cities; among these in attractiveness and the renown of their names Tyre, Sidon and Berytus are conspicuous, and equal to these are Emissa and Damascus, founded in days long past.) Loeb trans. Balty and Balty (1981) 60. 940 Discussed in detail on pp.239ff. 941 Purcell (1990). 273/448 all space is ideologically constructed, borderland spaces are particularly ripe for ambiguous and selective description between competing narratives. By their selection and presentation of material, use of historical comparanda, and treatment of imperial power and contemporary political relationships, the Roman imperial geographic writers created a geography of the Mesopotamian borderland centred on Roman power. In Chapter 2 I argued that the geographical writers presented local political structures in ambiguous terms. Their consideration of inter-imperial relations was the same strategy writ-large. In their perspectives on the Mesopotamian borderland, all non-Roman sources of contemporary authority were rendered vague and ambiguous, shifting and transient, and unchallenging to Roman dominion. All of these representations reveal the deliberate and selective nature of their geographic narratives, as well as contributing to our understanding of how Roman geographical authors conceived of the relationships between Rome, Parthia and the intervening borderland. 274/448 Chapter 5: Commercial Movement in the Borderland The definitive characteristic of a borderland are the boundaries which overlap and interact within it. Boundaries in turn are defined by how , where and why they attempt to prevent, restrict, and control movement. Movement and mobility thus lie at the heart of any examination of a borderland space, and the Mesopotamian borderland is no different. This chapter considers the presentation of commercial movement in Imperial Roman geographic writing by examining the productivity of northern Mesopotamia, the kinds of commercial activity and movement represented in the texts, and the interaction of that activity and movement with local and regional economies, communities and political structures. In considering the geographic presentation of movement, it is important to remember that people seldom universally conform to restrictions on movement. Where the possibility of movement exists someone will try to profit by it. Boundaries reflect and create differential conditions on either side, and such conditions provide opportunities for those willing to transgress those boundaries and create their own paths through space. 942 The geographical, political, cultural, demographic and economic boundaries of the Mesopotamian borderland afforded ample opportunity for profitable boundary crossing, most obviously involving the transport of goods, people and ideas between the Roman and Iranian spheres. By examining the kind of movement taking place and its presentation by the geographical writers, we can build a better picture of the interaction between those spaces and the spatial representation of the region. The geographical sources reveal movement in many ways: explicitly or implicitly; by inference from omissions or from narrative structure; by 942 Barth (2000); Sahlins (1989). The theoretical underpinnings of much work on the interactions between space, society and individual action lies in the work of Henri Lefebvre, in particular his La Production de l’espace (1974), translated into English in 1991: Lefebvre (1991); Soja (1996). 275/448 the presence of mobile people, mobile activities, or structures and places designed to serve mobile people and activities. While cognisant of the tendency for people to create their own paths through space, we may nevertheless examine certain frequently attested routes which attracted both the movement of considerable numbers of people, and the attention of the imperial geographic writers. The extant geographical narratives of Mesopotamia describe three main routes of inter-imperial border crossing, each of which is best conceptualised as a set of routes which pass through the same general areas. From north to south they are: 1) The northern route, which reached the Euphrates from the west through Roman Commagene or Syria and proceeded east across the rain-fed northern zone of Mesopotamia at the foot of the Taurus mountains towards Adiabene, the Tigris and Armenia; 2) the Euphrates route, which crossed the Euphrates in the same places, then more-or-less followed the river south-east towards the important cities of Parthian or Sasanian Babylonia; and 3) the Desert route, created and operated by Palmyrene caravans traversing the Syrian desert between Palmyra and various Palmyrene trading outposts in cities along the middle and lower Euphrates. The geographical writers also provide evidence of inter-provincial routes running north-south within the Roman empire as well as local routes. In this chapter, I will examine the presentation of each of these routes in the Roman imperial geographic sources. I begin with the Euphrates route, described in different ways by Strabo, Isidore and Ammianus. I then address the northern routes, alluded to by Pliny and Ammianus, but finding their fullest exploration in the Expositio Totius Mundi. Our main geographical source for the desert routes is Pliny , but a picture of the local movement in the Palmyrene hinterland emerges when Ptolemy’s Geography is examined in the light of the Palmyrene Tax Law . The final set of routes I consider are those which run north-south through the borderland. These are largely absent from the geographic narratives of the Mesopotamian borderland. The last two sections of the chapter are devoted to what little local movement is visible in these geographical narratives. Evidence for local routes appears mostly in the Antonine 276/448 Itineraries and the Peutinger Table, but the appearance of the Skenitai in geographical descriptions of the borderland provide an insight into another kind of movement. Overall, I argue that commercial and economic factors were a major influence on the way geographical sources constructed the borderland in nearly every case and that the Roman imperial geographic writers emphasise inter-Imperial connections over intra-Imperial and local links. Before proceeding with the analyses of these specific sets of routes, two matters require attention: northern Mesopotamia as a productive region in its own right and the context of long distance trade between the Roman empire and the regions to the east. These are both subjects in which the role of northern Mesopotamia is overshadowed by a greater degree of ancient evidence and contemporary scholarly interest in activity taking place in neighbouring regions, namely , the productivity of southern Mesopotamia/Babylonia and trade routes through the Red Sea to India. The Productivity of Mesopotamia Mesopotamia as a whole was a very productive region. Herodotus considered it even more so than the proverbially wealthy Egypt, even with the neighbouring regions included (Cyrene, Libya and Barca). 943 In Herodotus’ list of Persian tax regions, the tax paid by the ninth district (sometimes called the Great Satrapy of Mesopotamia) exceeded all other satrapies in the Persian Empire: 1000 talents of silver and 500 boy eunuchs. 944 Herodotus’ ninth district encompassed all of Mesopotamia from the Taurus to the Gulf of Arabia; the satrapies of Bâbiruš, Ahurâ and Arabâya. 945 While the intensely irrigated southern region of Babylonia may have produced higher yields per unit of land area which allowed the growth of great cities like Babylon, Seleucia, and Ctesiphon, northern Mesopotamia’s broad areas of rain-fed plains may have out- 943 Hdt. 3.92. 944 Hdt. 3.92: “Ἀπὸ Βαβυλῶνος δὲ καὶ τῆς λοιπῆς Ἀσσυρίης χίλιά οἱ προσήιε τάλαντα ἀργυρίου καὶ παῖδες ἐκτομίαι πεντακόσιοι· νομὸς εἴνατος οὗτος” (From the Babylonians and the rest of Assyria, 1000 talents of silver and 500 castrated boys; being the ninth district.) 945 Namely , southern Mesopotamia as formerly ruled from Babylon, northern Mesopotamia as formerly controlled by Assyria, and the steppe regions of the middle Euphrates: DB col.1, lines 14-15; DPe lines 10-11; DSe lines 25-26; DNa lines 26-27; DSAa; XPh lines 22 & 25; Briant (2002) 172–3; Herzfeld (1968) 292, 357–9. See Chapter 1. 277/448 produced the south. 946 Limits on the distance that farmers could transport their produce to markets and tax- collection centres meant that urban centres in the north had to be smaller and more dispersed than in southern Mesopotamia, but northern Mesopotamia was nevertheless able to support a number of large Bronze Age settlements. 947 Iron Age settlement of the region under the Assyrian empire was further dispersed among a larger number of smaller settlements, but the structures of that empire developed to draw produce from a wider area in support of the imperial cities on the banks of the Tigris. 948 The dispersed wealth and productivity of the northern area may have been less visible than the concentrated surpluses of Babylonia to the south, but when northern Mesopotamia was politically united under an empire, the capacity of that area would have been clear, at least to regional administrators. This may explain why the region’s fertility remains somewhat implicit in the Roman imperial authors. Our sources describe the fertility of the region in the Roman period. Both Strabo and Pliny refer to the fertility of the northern areas of the Mesopotamian borderland at the foot of the Taurus mountains; Strabo says that Mygdonia is εὐδαίμων ἱκανῶς, while Pliny calls it ubertatem. 949 Both authors begin their descriptions with these remarks, but do not return to the subject. Nevertheless, the prominent placement of 946 W eiss (1986); Wilkinson et al. (2005) esp. 484–85. Strabo (16.1.14) describes the produce of Babylonia. 947 Wilkinson et al. (1994); Ur (2003) 112. For pre-Hellenistic settlement history , see Liverani (1988); Fales (1990); Kühne (1995); Wilkinson (1994); Wilkinson (1995); Wilkinson (2000); Wilkinson (2005). On commercial activities in the Bronze Age, Lamberg-Karlovsky (1996) who notes on p.80: “There is no reason to disbelieve that once these [tax] quotas were achieved, the laborer was able to profit from his/her/their own private production. Everything produced beyond the quota could be freely disposed of in an open market of barter exchange. It is difficult to conceive of what other mechanism would have existed for satisfying the growing population’s everyday needs.” 948 In the 14 th century BCE, Assyrian control spread as far as the Euphrates and through the subsequent centuries coalesced through the growth of agricultural colonisation, networks of communication and transports, and strongpoints for defence and tribute collection. Liverani (1988) esp. pp. 90–92; Wilkinson (2005) 37–44. Smaller Bronze Age sites would not have needed extensive inter-regional networks to supply their basic needs, but the larger size of Assyrian capitals would have required the use of long distance networks of up to 100km just to supply the agricultural needs of the urban population: Wilkinson (2005) 36; Altaweel (2004). Furthermore, lower yields per area reduce the importance of particularly fertile patches of land and has and evening effect on the distribution of wealth. Weiss argues that competitive or adaptive advantages of absorb climatic uncertainty through land-use strategies produced a more even playing field in northern Mesopotamia. W eiss (1986) 98. 949 Strabo 16.1.23; Pliny 6.117. I take the description of extreme fertility at Strabo 16.1.14 to refer to Babylonia specifically rather than the entire land of the Assyrians. 278/448 the statements suggests that fertility was an important characteristic of northern Mesopotamia. Historical descriptions of events in the region bear this out. Describing a gift of part of Armenia to a Parthian king, Josephus describes the country around Nisibis as ἀγαθός, surely a reference to its fertility . 950 In Polybius' account of Antiochus III's march from Antioch against the rebellious satrap Molon, Antiochus and his army sat out the worst of the winter for 40 days at Nisibis. 951 This implies the ability to store substantial quantities of food drawn from a fertile hinterland. In Plutarch’s account of Lucullus’ campaign against Tigranes in the first century CE, Lucullus had defeated the Armenian forces blocking the road to their capital at Artaxata (near modern Artashat), but bad weather and discontent in the ranks forced him to turn south. The army descended into Mygdonia, a χώραν παμφόρον καὶ ἀλεεινὴν (“an area productive and open to the sun”). 952 In his Life of Apollonius, Philostratus’s description of Mesopotamia as a land of villages reflects the archaeological evidence. 953 Ammianus Marcellinus’ geographical description of northern Mesopotamia is lost, but he refers to the region’s fertility in the course of describing historical events, such as the return of Julian’s army to Roman territory . 954 Military expeditions down the Euphrates often returned to Roman territory through the fertile northern part of the borderland, a tendency which ancient authors often 950 Josephus Ant. Iud. 20.68: “ἔδωκεν δὲ καὶ χώραν πολλὴν αὐτῷ κἀγαθὴν τοῦ τῶν Ἀρμενίων βασιλέως. Νίσιβις δέ ἐστιν ὄνομα τῇ γῇ, καὶ ἐν αὐτῇ πρότερον Μακεδόνες ἐκτίσαντο πόλιν Ἀντιόχειαν, ἣν Ἐπιμυγδονίαν προσηγόρευσαν.” (He furthermore gave him an extensive and productive territory which he carved from that of the king of Armenia. The district is called Nisibis, and in it the Macedonians had in days of old founded the city of Antioch which they surnamed Epimygdonia.) Loeb trans. Josephus is confused; Nisibis is a city not a district, and is the same place as Antioch Epimygdonia, Mygdonian Antioch. Cassius Dio (68.26.1) mentions the forests around Nisibis which existed in Trajan’s time. Ammianus Marcellinus (19.9.5) also mentions forests between Samosata and Nisibis in the fourth century. 951 Polybius 5.51.1. 952 Plut. Luc. 32.4: “ὡς δ’ οὐκ ἔπειθεν, ἀπῆγεν αὐτοὺς ὀπίσω, καὶ κατ’ ἄλλας ὑπερβολὰς διελθὼν τὸν Ταῦρον, εἰς τὴν λεγομένην Μυγδονικὴν κατέβαινε, χώραν παμφόρον καὶ ἀλεεινὴν καὶ πόλιν ἐν αὑτῇ μεγάλην καὶ πολυάνθρωπον ἔχουσαν, ἣν οἱ μὲν βάρβαροι Νίσιβιν, οἱ δ’ Ἕλληνες Ἀντιόχειαν Μυγδονικὴν προσηγόρευον.” (But since he could not persuade them, he led them back, and crossing the Taurus by another pass, descended into the country called Mygdonia, which is fertile and open to the sun, and contains a large and populous city , called Nisibis by the Barbarians, Antioch in Mygdonia by the Greeks.) Loeb trans. 953 Philostratus, Apollonius 1.20.2: “...εν ᾗ και πολεις μεν, τὸ δὲ πλεῖστον κῶμαι...” (...in which are some cities but mostly villages...). 954 Amm. Marc. 23.6.15-25. While his description of Assyria refers to the great wealth of that region (ditissima), his description of that region concerns Adiabene, Babylonia and Mesene. 279/448 attributed to logistical concerns. 955 The ample opportunities for supply in the north were regularly contrasted with scarcity and lack in the harsh desert environment of the lands along the Euphrates. The Roman imperial geographic writers show the Mesopotamian borderland as a fertile and productive region, yet they seldom show the activity of local networks or its participation in intra-imperial networks of exchange and commercial movement. As well as indicating the production of agricultural wealth directly , this fertility and imperial control imply the existence of local distribution networks to transfer foodstuffs and taxation to the cities and goods produced and procured in the urban areas to the countryside. There is little explicit discussion of these, but they can be viewed implicitly . In this chapter I show how the Roman Imperial geographic writers focus their attention on inter-imperial routes at the expense of these local and intra-imperial routes. As background for these arguments, an overview of Roman long-distance trade in the east is in order. Long-distance trade What was the role of the Mesopotamian borderland in eastern trade? Roman trade with the states and people to the east seems to have been along two main vectors: by sea to India and overland through the Parthian empire via “the silk road” to Central Asia and China. Once again, these two routes actually reflect several paths which could be taken by enterprising merchants, often combining land and sea travel. Several recent works have analysed trade interactions between Rome and the East; it is not necessary to re-examine their arguments in detail here. Rather, I will briefly summarise the main points salient to an examination of the commercial networks visible in descriptions of the Mesopotamian borderland. 956 955 Supply difficulties are often cited explicitly as the reason: Amm. Marc. 24.7.6 (Julian); Dio 76.9.3-5 (Septimius Severus). Trajan returned via Hatra: Dio 68.31. Nevertheless, the frequent use of the Euphrates route by large armies suggests that opportunities for water-borne supply mitigated any deficiency of forage and that a certain degree of rhetorical exaggeration may have been at work in our sources. 956 Further detail on trade between Rome, India, Arabia and Iran can be found in Young (2001); Tomber (2008); Morley (2007); Parker (2008); Casson (1989); Parker (2001); Salles (1998). A brief bibliography of older works can be found in Smith (2013) 189n5. 280/448 The goods which were the object of these long distance routes were high-value and portable items like silk and spices. Grant Parker’s examination of a third century list of goods subject to tax at Alexandria (the sole Roman customs point for long-distance trade through Egypt) shows that items with medicinal, culinary , cosmetic, aromatic and decorative uses were transported through the Red Sea. 957 Although it has been, and remains, common to describe these as “luxury goods”, that term is anachronistic and fails to account for the use of these items in religious and medical practices. 958 The common factor is not a subjective category like “luxury”, but high degree of value for their weight and bulk. Given the high risk and cost of long-distance travel by land or sea, the opportunity for profit must have been considerable. Nevertheless, the volume of trade is difficult to estimate. 959 Commercial links through the Red Sea to India preceded Roman control of Egypt but accelerated under the Julio-Claudians. 960 Ptolemaic kings established infrastructure linking the Red Sea to the Nile valley and appear to have played a quite active role in developing the route. 961 Initially Greek Merchants only operated in the Red Sea itself, trading with Indian merchants at Aden. Around the end of the second century BCE, Greek sailors figured out how to sail with the monsoon and began to ply the entire route between Egypt and India. 962 Archaeological research has revealed substantial quantities of Roman material at Indian sites, especially from the second century CE, as well as imperial investment in the development and maintenance of Red Sea and Nile ports and the road and canal system between them. 963 The trade with India 957 Parker (2002) 41–55. On goods mentioned by the Periplus Maris Erythraei specifically: Casson (1989) 21–27. 958 Young (2001) 14–18; Tomber (2008) 15–16 (with bibliography). 959 Young (2001) 24–26. Whittaker 1998, p.20 “To reach out to india and pursue the dawn:” the Roman view of India, Studies in History 14.1:1-20. Caution on extent: Parker 2002, pp.74-75; danger of modern comparison. And of attempts to use statistics reported in ancient works, Parker 2002 pp.77-78. 960 On commercial links between Roman Egypt and India, see Young (2001) 27–89; Tomber (2008); Salles (1998); Casson (1989). (Tomber (2008) 57–87) 961 Young (2001) 27; Parker (2001) 73; Tomber (2008) 60–64; Parker (2002) 69; Casson (1989) 36. 962 Casson (1989) 12; Young (2001) 28–32. Trade with Arabia continued as a separate undertaking, governed by the timing of the insense harvest rather than the monsoon season, Young (2001) 34–36. 963 Indian sites: Casson (1989) 12; Tomber (2008). Egyptian infrastructure: Casson (1989) 13–14; Young (2001) 38– 45 (ports), 45–54 (Julio–Claudian and Flavian roads), 74–79 (second century roads and canals). 281/448 was of such a volume that a substantial and permanent Roman community seems to have been established in the southern Indian city of Muziris. 964 Roman coins were exchanged as far as Sri Lanka and the geographers included descriptions of that island, no doubt informed by travellers reports. 965 Numismatic evidence suggests the trade was at its peak under the Julio-Claudians and began to decline in the second century , with brief revivals during the reigns of Antoninus Pius and Septimius Severus. 966 The disrupted conditions of the third century Roman Empire seem to have reduced Roman participation in the Red Sea routes and allowed the rising power of the Axumites to the south of the Red Sea to begin to dominate the route to India. 967 Nevertheless, the fortunes of Roman merchants revived in the Diocletianic period and remained active at least into the fifth century CE. 968 Before and during the first century CE, this pattern of development and activity is borne out by the literary sources. Artemidorus’ description of the Hellenistic route along the Red Sea coast is relayed to us by Strabo. 969 Despite the likelihood that such coastal periploi derived from mercantile movement or were at least reliant on sailing expertise based on that activity , Strabo is dismissive of the reliability of private citizens engaged in commercial practices. 970 However, he notes that the growth of such trade was a beneficial outcome of the Roman annexation of Egypt, boasting on two occasions that since the institution of Roman rule, large fleets had begun to work the routes between Egypt and India, with a specific emphasis on the 964 Young (2001) 30–31; Documentary evidence for “Yavana” (Greeks) at Muziris: Parker (2002) 63–64. 965 Strabo (15.1.14-15) quotes Eratosthenes, Onesicritus and mentions that other authors described the island. Pliny 6.89 on trade with Taprobane. Taprobane is the last region to be described in Ptolemy’s work (7.4); it lies narratively at the edge of the world. Young (2001) 32–33; Parker (2002) 78; Walburg (2008) 319–42. 966 Young (2001) 74–79; Tomber (2008) 30–37. 967 Parker (2002) 71–72. Mlasowsky and Pahlitzsch, " Axum, Axomis." BNP . 968 Young (2001) 86–88. 969 Strabo 16.4.5ff. 970 Strabo 15.1.4: “Καὶ οἱ νῦν δὲ ἐξ Αἰγύπτου πλέοντες ἐμπορικοὶ τῷ Νείλῳ καὶ τῷ Ἀραβίῳ κόλπῳ μέχρι τῆς Ἰνδικῆς σπάνιοι μὲν καὶ περιπεπλεύκασι μέχρι τοῦ Γάγγου, καὶ οὗτοι δ’ ἰδιῶται καὶ οὐδὲν πρὸς ἱστορίαν τῶν τόπων χρήσιμοι.” (As for the merchants who now sail from Egypt by the Nile and the Arabian Gulf as far as India, only a small number have sailed as far as the Ganges; and even these are merely private citizens and of no use as regards the history of the places they have seen.) Loeb trans. By “Arabian Gulf”, Strabo probably means the Red Sea. Strabo 17.1.13. 282/448 collection of taxes on both incoming and outgoing cargoes. 971 In these boasts, Strabo reveals the considerable expansion of the route under Augustus. Towards the end of the Julio-Claudian period, the Periplus Maris Erythraei provides our best source for the operation of this route. This anonymous Greek text, written around 40-70 CE, probably by an businessman based in Egypt, attests to a thriving commercial link between Roman Egypt and the cities and states of southern Arabia and western India at that time. 972 Soon after, Pliny describes the route between Alexandria and India, specifying the dates of the important monsoon wind pattern in the northwest Indian Ocean. 973 Pliny’s report that reliable information on the geography of South Asia has only recently become available points to the growth of this trade route, and to the production of texts like the Periplus Maris Erythraei in the Flavian period. 974 Although seasonal prevailing winds hinder or prevent certain sailing routes, this maritime route around Arabia was the main channel for Indian goods to arrive in the Roman Empire. The narrative of the Periplus Maris Erythraei skips across the mouth of the Persian Gulf, but notes that exports from the Gulf could be exchanged in the cities on either side of its mouth. 975 There is little evidence of Roman commercial activity in the Gulf itself; it seems that it played a minor role in Indo-Roman trade. 976 The route around Arabia described by the Periplus Maris Erythraei was the major connection between the Roman Mediterranean and India; most such trade bypassed the overland routes through Babylonia and Mesopotamia. 971 Strabo 2.5.12; 17.1.13. See also his remarks on the important Egyptian ports of Berenike and Myus Hormus: Strabo 17.1.45. 972 Casson (1989) 6–10; Parker (2002) 62–63. Casson notes several first person references which suggest the author considered Egypt his home. 973 Pliny NH 6.102-104. Casson (1989) 13–15. On Pliny and trade generally , see French (1994) 210–12. 974 Pliny 6.101; Parker (2001) 71. 975 Casson (1989) 19. 976 Tomber (2008) 109–16; Rutten (2007) 17–19. There is evidence of trade between Babylonia and India from the Bronze Age to the Seleucid period (Lamberg-Karlovsky (1996) 87–88; Parker (2002) 70). The Achaemenids and Seleucids maintained a navy in the Arabian Gulf to support their trading ports. The latter empire may have pursued a deliberate policy with commercial objectives: Salles (1987) esp. 88–99. This trade continued in the Parthian and Sasanian periods, although evidence of Roman trade far outweighs the Mesopotamian material (Tomber (2008) 79–80, 99–100, 146–47). Most of this evidence comes from northwest India and consists of storage vessels, raising the questions of whether Roman goods were traded and repackaged in Babylonia, of the visibility of goods not requiring permanent vessels, and of the nature of the trade relationships between northwest India and the eastern regions of the Parthian and Sasanian empires. 283/448 It used to be common to speak of the Parthians controlling and monopolising overland trade, especially the “silk road” to central Asia and China. 977 However, India may have been the Roman Empire’s main source of silk. 978 Certainly there is no evidence of intentional manipulation or obstruction of overland trade by the Parthians. 979 In fact, their permissiveness regarding the operation of Palmyrene merchants in their territory suggests, if anything, the opposite. 980 The Roman preference for the Red Sea route was probably a matter of convenience and the relative efficiency of sea transport. 981 However, the importance of that route did not render other paths redundant, especially those which allowed access to inland areas. Strabo and the Euphrates Route Strabo provides one of our best sources for movement in the Mesopotamian borderland, especially for mercantile activity along the Euphrates. Strabo provides a concrete and personal description of this route entwined with a mercantile perspective on matters such as taxation and the logistics of travel through arid areas. However, as I will show , his description treats the route as passing through an ambiguous space between empires. This differs from the accounts of Isidore and Ammianus who describe the route as a topologically located pathway connecting settlements within the space. In his description of Mesopotamia, Strabo describes a route beginning in Syria which, except for a brief diversion through the desert, proceeds down the Euphrates to a place called Skenae on the borders of Babylonia (Map 24: Strabo's Euphrates Route). διὰ δὲ τῶν Σκηνιτῶν, ὑπὸ τῶν Μαλίων νυνὶ λεγομένων, καὶ τῆς ἐκείνων ἐρημίας ἡ ὁδὸς τοῖς ἐκ τῆς Συρίας εἰς Σελεύκειαν καὶ Βαβυλῶνα ἐμπορευομένοις ἐστίν. ἡ μὲν οὖν διάβασις τοῦ Εὐφράτου κατὰ τὴν Ἀνθεμουσίαν ἐστὶν αὐτοῖς, τόπον τῆς Μεσοποταμίας· ὑπέρκειται δὲ τοῦ 977 Schoff (1989) 18–19; Warmington (1974) 20; Miller (1969) 139, 235; Thorley (1971) 75. For a thorough recent overview of the history and archaeology of the Silk Road, see Hansen (2012). 978 Young (2001) 28; Raschke (1978) 630–31. 979 Young (2001) 195–98. 980 See below , pp.308ff. 981 In the 19 th century , British experiments with an overland route between the Mediterranean and India along the Euphrates illustrated several problems with riverine transport through a borderland region. See Guest (1992). 284/448 ποταμοῦ σχοίνους τέτταρας διέχουσα ἡ Βαμβύκη, ἣν καὶ Ἔδεσσαν καὶ Ἱερὰν πόλιν καλοῦσιν, ἐν ᾗ τιμῶσι τὴν Συρίαν θεὸν τὴν Ἀταργάτιν. διαβάντων δὲ ἡ ὁδός ἐστι διὰ τῆς ἐρήμου ἐπὶ τοὺς τῆς Βαβυλωνίας ὅρους μέχρι Σκηνῶν, ἀξιολόγου πόλεως ἐπί τινος διώρυγος ἱδρυμένης. ἔστι δ’ ἀπὸ τῆς διαβάσεως μέχρι Σκηνῶν ἡμερῶν πέντε καὶ εἴκοσιν ὁδός. καμηλῖται δ’ εἰσί, καταγωγὰς ἔχοντες τοτὲ μὲν ὑδρείων εὐπόρους τῶν λακκαίων τὸ πλέον, τοτὲ δ’ ἐπακτοῖς χρώμενοι τοῖς ὕδασι. παρέχουσι δ’ αὐτοῖς οἱ Σκηνῖται τήν τε εἰρήνην καὶ τὴν μετριότητα τῆς τῶν τελῶν πράξεως, ἧς χάριν φεύγοντες τὴν παραποταμίαν διὰ τῆς ἐρήμου παραβάλλονται, καταλιπόντες ἐν δεξιᾷ τὸν ποταμὸν ἡμερῶν σχεδόν τι τριῶν ὁδόν. οἱ γὰρ παροικοῦντες ἑκατέρωθεν τὸν ποταμὸν φύλαρχοι, χώραν οὐκ εὔπορον ἔχοντες, ἧττον δὲ ἄπορον νεμόμενοι, δυναστείαν ἕκαστος ἰδίᾳ περιβεβλημένος ἴδιον καὶ τελώνιον ἔχει, καὶ τοῦτ’ οὐ μέτριον. χαλεπὸν γὰρ ἐν τοῖς τοσούτοις καὶ τούτοις αὐθάδεσι κοινὸν ἀφορισθῆναι μέτρον τὸ τῷ ἐμπόρῳ λυσιτελές. διέχουσι δὲ τῆς Σελευκείας αἱ Σκηναὶ σχοίνους ὀκτωκαίδεκα. 982 982 Strabo 16.1.27: “The route for those travelling as merchants from Syria to Seleuceia and Babylon runs through the country of the Skenitai, now called Malians by some writers, and through their desert. Such travellers cross the Euphrates near Anthemusia, a place in Mesopotamia; and above the river, at a distance of four schoinoi, lies Bambycê, which is also called Edessa and Hierapolis, where the Syrian goddess Atargatis is worshipped; for after they cross the river, the road runs through the desert to Skenai, a noteworthy city situated on a canal towards the borders of Babylonia. The journey (ὁδός) from the crossing [of the Euphrates] to Skenai takes twenty-five days. On it there are camel-drivers who maintain stopping places which are sometimes well-supplied with water from cisterns, but at other times using imported water. The Skenitai are peaceful and moderate towards these travellers in exacting tolls, so those travellers risk travel through the desert in order to avoid the lands near the river ( τὴν παραποταμίαν), leaving the river on the right for a journey of almost three days. For the chiefs of those living on both sides of the river, holding land which is not rich, but is less lacking than that of some others, is each sovereign in his particular lands and charges tolls, and these are not moderate. For this reason it is difficult among so many and such stubborn people for a common measure (of taxation) to be set which benefits the merchant. Skenai is eighteen schoinoi distant from Seleuceia.” The locations mentioned in this passage are discussed in Chapter 1 and Appendix 1. 285/448 The length and detail of this passage attests to its importance in Strabo’s conception of Mesopotamia. 983 Strabo’s description of the topology of the route itself is vague, but what his description lacks in geographical specificity , it makes up for in the description of the activities facilitated by the route. This route (ὁδὸς) crosses the Euphrates near Anthemusia (κατὰ τὴν Ἀνθεμουσίαν) and travels through the southern desert regions (διὰ τῆς ἐρήμου) for 25 days, mostly following the Euphrates, until it reaches Skenai, ἀξιολόγου πόλεως ἐπί τινος διώρυγος ἱδρυμένης (“a noteworthy city situated on a canal”) in Babylonia, probably to be identified with Misiche on the Naarmalcha canal. 984 Rather than give a detailed account of the 983 This passage comprises the bulk of Strabo’s description of southern Mesopotamia. 984 The combination of the imprecision of written sources and difficulties of interpreting the archaeological remains of paths which often remain in use today means that the question of routes through this region is a matter of debate. On the various suggestions for crossing points on the Euphrates between northern Syria (including Commagene) and Osrhoene, from Assyrian times, see: Gawlikowski (1996); Comfort et al. (2000); Comfort and 286/448 Map 24: Strabo's Euphrates Route course of the route or the settlements it passed through, Strabo describes a few cities along the way and generally refers to the amenities available to merchants during the journey . Strabo describes stations and water supplies, alludes to the collection of taxation by both the Skenitai and the settlements along the river itself. Strabo’s ὁδὸς is more than an abstract and impersonal link between two spaces. The path which the route takes is not defined topologically , but the distance of the journey is specified in units of travel time (days and skoinoi). 985 This embeds personal mobility in the language of the passage. Furthermore, throughout the passage, Strabo implicitly defines the direction of the route as leading from Roman to Parthian space, placing the agency for the movement firmly in Roman hands. The people moving along the route are described as ἐμπορευομένοις, “those travelling as merchants”. In Strabo’s presentation, this is a mercantile route used by traders between the Roman and Parthian empires. Strabo further highlights the role of commercial movement by framing his description of the route with the perspective of travelling and trading merchants. He introduces his description of the route with a sentence placing the movement of merchants through the lands of the Skenitai (διὰ δὲ τῶν Σκηνιτῶν... καὶ τῆς ἐκείνων ἐρημίας ἡ ὁδὸς τοῖς... ἐμπορευομένοις ἐστίν). 986 Only after couching the passage in such terms does he proceed with what geographical detail he gives. At the end of the passage, after describing the tax-collection policies of the phylarchs along the river, Strabo gives an aside lamenting the difficulties which those policies impose upon the merchant: χαλεπὸν γὰρ ἐν τοῖς τοσούτοις καὶ τούτοις αὐθάδεσι κοινὸν ἀφορισθῆναι μέτρον τὸ τῷ ἐμπόρῳ λυσιτελές. 987 Except for a final, brief geographical note on the distance between Skenai and Ergeç (2001). For the identification of Skenae with Misiche, see Chapter 1. For more on Misiche and the Naarmalcha canal, see Appendix 1.12. 985 There were two methods available for relating distance, travel time or absolute distance using the stadion. 986 Strabo 16.1.27: “The route for those travelling as merchants runs through the Skenitai and their desert”. 987 Strabo 16.1.27: “ For this reason it is difficult among so many and such stubborn people for a common measure (of taxation) to be set which benefits the merchant” Strabo’s use of ἀφορίζω (to mark off by boundaries; ὅρος, boundary) is also interesting in this passage. 287/448 Seleucia, the passage begins with the movement of merchants and ends with this editorial comment from the narrator on the difficulty of gaining commercial advantage in that profession. Other mobile parties support the movement of merchants in Strabo’s narrative. The most important of these groups were the καμηλῖται, camel-guides, who maintained stopping places (καταγωγὰς) along the route. Strabo’s description alludes to a high degree of organisation by these camel-guides in conjunction with local Skenitai tribes. The camel-guides ensured that these stopping points were supplied with water drawn from cisterns (τῶν λακκαίων) or imported (ἐπακτοῖς τοῖς ὕδασι). He is not specific about the sources or the mechanism by which this supply occurs, but several possibilities exist. The cisterns could either have been natural wells or oases, or they could have been constructed to collect and store rainwater. 988 Rain is scarce in the Jazira and while it can be heavy when it does occur, it evaporates quickly . Evidence of networks of shallow channels and artificial cisterns for the collection of rainwater as been found in areas where such water supplies might be required, such as the Syrian desert and the Sinai. 989 Strabo states that the stopping places were supplied with water, not that the caravans transported water with them. This implies an organisation of water-transportation to supply the cisterns along the route. This imported water was presumably transported to the stopping places in skins tied to the backs of camels. This suggests the circulation of information about cistern levels among camel-guides and their water-transporting agents; depending on the relative size of the cisterns and the caravans, the process may have required timely information to ensure the caravans were adequately supplied. The camel-guides probably also guarded travellers from hostile Skenitai tribes; Strabo notes that the desert was risky for merchants (διὰ τῆς ἐρήμου παραβάλλονται). The circulation of information among local networks probably assisted this protective role. The καμηλῖται which Strabo describes probably made use of various personal, family and economic 988 Hoyland (2001) 85. Crossing the Sinai peninsula required similar logistical considerations, Cruz-Uribe (2003) 22–24. 989 Kennedy and Riley (1990) 70–76; Cruz-Uribe (2003) 23–24. 288/448 ties to the nomadic pastoralists through which they passed to facilitate the passage of goods down the Euphrates, as did the Palmyrene caravans to the south. 990 Another major interaction between travellers and local groups was taxation. According to Strabo, the Skenitai of the desert were moderate in their tolls (τὴν μετριότητα τῆς τῶν τελῶν πράξεως) while the phylarchs of the river communities were not (τελώνιον ἔχει, καὶ τοῦτ’ οὐ μέτριον). To avoid the taxation of those phylarchs, some travellers leave the river for almost three days and pass through the territory of Skenitai groups. 991 Strabo characterises this desert journey as risky (ἧς χάριν φεύγοντες τὴν παραποταμίαν διὰ τῆς ἐρήμου παραβάλλονται), presumably in part because not all of the Skenitai were peaceful. However, while Strabo implies that all merchants took the desert route, he also implies that accepting the taxation of the phylarchs near the river was an option. It seems that, the cost of hiring καμηλῖται and caravan guards for a desert journey was less of a burden than the taxation imposed while following the route down the river. 992 If the desert route was cheaper, did every merchant who could secure καμηλῖται avoid the river? Probably not. While Strabo does not draw out the specific economic implications, merchants had a choice between the higher-risk desert route with lower fixed costs and a lower-risk route along the river with higher fixed costs. If Strabo’s characterisation of the high taxes set by the river φύλαρχοι is correct, that desert 990 The success of Palmyrene caravan leaders at managing their desert routes and the tribes through which they passed elevated Palmyra to one of the wealthiest and most important cities of Syria. See chapters 3 and 6. 991 Strabo 16.1.27 “καταλιπόντες ἐν δεξιᾷ τὸν ποταμὸν ἡμερῶν σχεδόν τι τριῶν ὁδόν”, (leaving the river on the right for a journey of almost three days). It is unclear whether Strabo refers to a three day journey or a longer journey , running parallel with the Euphrates three days to the north. Strabo characterises the desert diversion as an attempt to avoid taxation by local rulers along the river, but it may simply refer to the initial stages of Isidore’s route across the arid steppe between Zeugma and the Balikh river: Tarn, 'Parthia' CAH 9 (1963), p.608 identifies the route taken by Crassus (Plut Cras 22) as the route described by Strabo (16.1.27); Millar (1993) 440. If the route described be Strabo did run from Zeugma to the Balikh, it may have passed through Edessa, Ross (2001) 16. 992 Strabo 16.1.27: “οἱ γὰρ παροικοῦντες ἑκατέρωθεν τὸν ποταμὸν φύλαρχοι, χώραν οὐκ εὔπορον ἔχοντες, ἧττον δὲ ἄπορον νεμόμενοι, δυναστείαν ἕκαστος ἰδίᾳ περιβεβλημνος ἴδιον καὶ τελώνιον ἔχει, καὶ τοῦτ’ οὐ μέτριον. χαλεπὸν γὰρ ἐν τοῖς τοσούτοις καὶ τούτοις αὐθάδεσι κοινὸν ἀφορισθῆναι μέτρον τὸ τῷ ἐμπόρῳ λυσιτελές” (For the chiefs of those living on both sides of the river, holding land which is not rich, but is less lacking than that of some others, is each sovereign in his particular lands and charges tolls, and these are not moderate. For this reason it is difficult among so many and such stubborn people for a common measure (of taxation) to be set which benefits the merchant.) 289/448 journey was perceived (at least by Strabo) as quite risky indeed. The region’s nomadic pastoralists are central to Strabo’s description of the route. While Strabo does not provide topographical markers for most of the journey , the space through which the movement passes is defined in both ethnic and geographical terms: the route passes through the Skenitai (διὰ δὲ τῶν Σκηνιτῶν) and their desert (καὶ τῆς ἐκείνων ἐρημίας). Their role in maintaining the route which allowed mercantile movement down the Euphrates has been described, their second role as an obstacle, or at least a friction, to movement has only been alluded to. Although in this section, Strabo declares that the Skenitai are peaceful, elsewhere he refers to their practice of banditry . Strabo uses the term “Skenitai” broadly of nomadic pastoralists in the arid regions of Syria and Mesopotamia, ignoring or unaware of different cultural or tribal groups that certainly existed under that blanket term. The existence of these differences can be seen when Strabo gives contradictory information about the Skenitai, especially when he refers to Skenitai within the same geographical area. This is the case for the Skenitai of the Jazira. In his general description of the Jazira, just before his description of the Euphrates route in which Skenitai are peaceful assistants to travelling merchants, he describes them as wide-ranging bandits (λῃστρικοί). 993 When Strabo characterises the desert journey as risky (ἧς χάριν φεύγοντες τὴν παραποταμίαν διὰ τῆς ἐρήμου παραβάλλονται), he is probably thinking of both the environmental dangers of desert travel and the risks of banditry by the Skenitai who occupied the land through which the movement passed. Most likely , travellers employed camel-guides who were able to facilitate payment to certain Skenitai groups and ward off groups who refused or who were not offered payment. 993 Strabo 16.1.26: “τὰ δὲ πρὸς μεσημβρίαν κεκλιμένα τῆς Μεσοποταμίας καὶ ἀπωτέρω τῶν ὀρῶν ἄνυδρα καὶ λυπρὰ ὄντα ἔχουσιν οἱ σκηνῖται Ἄραβες, λῃστρικοί τινες καὶ ποιμενικοί, μεθιστάμενοι ῥᾳδίως εἰς ἄλλους τόπους, ὅταν ἐπιλείπωσιν αἱ νομαὶ καὶ αἱ λεηλασίαι. τοῖς οὖν παρορείοις ὑπό τε τούτων κακοῦσθαι συμβαίνει καὶ ὑπὸ τῶν Ἀρμενίων· ὑπέρκεινται δὲ καὶ καταδυναστεύουσι διὰ τὴν ἰσχύν.” (The Skenitai Arabs occupy those parts of Mesopotamia which incline towards the south and are farther from the mountains, and are waterless and barren. [The Skenitai] are bandits and shepherds, who readily move from place to place when pasture and booty fail them. As a result, the people near the mountains are harassed by the Skenitai and also by the Armenians, who dwell above them and oppress them through brute force.) 290/448 The centrality of the mercantile perspective to the route Strabo describes highlights the role of the route in Strabo’s conception of Mesopotamia as a whole. It is of note that Strabo’s description of the route largely omits any specific description of settled urban areas. He locates the Euphrates crossing with respect to Anthemusia and adds the specific comment that he is referring to a city on the east side of the river (τόπον τῆς Μεσοποταμίας). His confused conflation of Bambyce/Hierapolis and Edessa may indicate that both cities were near the route he describes. 994 Strabo either selectively omits the important Hellenistic settlements of Nikephorion, Circesium and Dura Europus, or conceals them under the rubric of “the phylarchs along the river” (οἱ παροικοῦντες ἑκατέρωθεν τὸν ποταμὸν φύλαρχοι). Thus the route is detached from the major urban centres on the banks of the Euphrates. Moreover, in Strabo’s account, this mercantile activity is quite unconnected to the sedentary culture of the fertile northern areas of Mesopotamia. The merchants travelling along the Euphrates only interact with the settled communities along the river which tax passing caravans and with the mobile communities of Skenitai who facilitate Mesopotamian transshipment by serving as guides and guards and exploit it by preying on merchants. Moreover, in Strabo’s description, the φύλαρχοι who interact with and benefit from the travelling merchants are only concerned with the tax revenue they can gather from the passing traders, not with any other benefits that might accrue from their presence. Their interaction is solely described in terms relevant to the route as a transshipment route through Mesopotamia. This transshipment route is the major link Strabo draws between Mesopotamia and empires to the east and west. In fact in Strabo’s presentation, the route does not serve to connect northern Mesopotamian cities to each other and to the outside world at all, but instead acts as a conduit between Mediterranean and Iranian space. For Strabo, this transit is the only local economic activity that extends beyond Mesopotamia itself, making Mesopotamia a transitory space of movement between the Roman and Iranian empires. The route is more of a highway through Mesopotamia than part of an integrated regional network. It is a transitory space 994 See Chapter 1. Bambyke/Hierapolis and Edessa are discussed further in Appendices 1.4 and 1.10. 291/448 through which connections between the Roman and Parthian empires flow , but in which the urban centres of Mesopotamia do not participate. Strabo’s description of the movement of merchants along the Euphrates illustrates several ways that the sources reveal physical mobility . The first is through explicit notifications of movement, such as the route from Syria to Babylonia used by merchants (ἐμπορευομένοις), as well as other groups who move along with them such as camel-guides (καμηλῖται). More common are implicit notifications, such as the existence of the reverse route, from Babylonia to Syria. Strabo’s narrative focuses exclusively on the Roman perspective and thus highlights Roman agency in trans-Mesopotamian movement. However, people and goods move in both directions and use routes in oblique ways not seen or foreseen by those who create and describe the routes. 995 Another kind of implicit description is that conveyed by the construction of the narrative itself. Descriptions of space based on movement, such as itineraries or periploi, can often be seen within works of descriptive geography , even those that initially seem to take a broader, static and panoptic view . 996 Strabo’s descriptions often follow lines of communications such as those afforded by rivers, roads or coastlines. 997 His inclusion of Edessa in this description of the Euphrates route implies that Edessa was a feature on that route, as it was on the Euphrates routes described by other authors. Two other geographic sources describe the Euphrates route between Syria and Babylon in greater topographic detail than Strabo: both Isidore of Charax and Ammianus Marcellinus describe a series of locations along the Euphrates between the Euphrates near Zeugma and Babylonia. 998 These three authors each describe and characterise the route in a different way and choose to include different sites, but the river provides a fixed physical feature from which those using the route seldom deviated. 999 995 Barth (2000). 996 See Introduction. Dueck (2000) 40–43. 997 Dueck (2000) 167–8. 998 Isidore 1; Amm. Marc. 23.2-24.6. 999 Although the route followed the Euphrates river, the journey does not seem to be conducted by boats, except for crossings. Prior to twentieth-century irrigation and hydro-electric projects, which regulate the flow of the river, 292/448 The Parthian Stations (Stathmoi Parthikoi) of Isidore of Charax is probably an excerpt from a longer work on Parthia written in the first century BCE. 1000 The Parthian Stations describes a route from the crossing of the Euphrates at Zeugma to Alexandria in Arachosia, near modern Kandahar in southern Afganistan. Müller suggested that Isidore is to be equated with the Dionysius of Charax, whom Pliny relates was sent to the east to research a work on the Parthians for Gaius Caesar: Hoc in loco genitum esse Dionysium, terrarum orbis situs recentissimum auctorem, quem ad commentanda omnia in orientem praemiserit Divus Augustus ituro in Armeniam ad Parthicas Arabicasque res maiore filio, non me praeterit nec sum oblitus sui quemque situs diligentissimum auctorem visum nobis introitu operis: in hac tamen parte arma Romana sequi placet nobis Iubamque Regem, ad eundem Gaium Caesarem scriptis voluminibus de eadem expeditione Arabica. 1001 As Chaumont has pointed out, while it is possible that the Isidore of the Parthian Stations and Pliny’s Dionysius are the same person, there is no requirement that they be so, nor is there a problem assuming they the level of Euphrates was inconsistent and sandbars and flood debris were a hazard. Moreover, while the trip down river may have been relatively easy , the journey back against the flow was problematic for nineteenth century steamboats, let along simpler forms of transport, Guest (1992). Herodotus (1.194) describes Armenians descending the Tigris to Babylon in boats made of skins and laden with goods for sale and donkeys for the return trip. The boatmen would sell the goats and the skins and return to Armenia on the donkeys. The passage is discussed in the context of Assyrian use of both the Tigris and Euphrates for transportation by Fales (1995) 211–15. Briant also discusses the passage, but transposes it from the Tigris to the Euphrates; Briant (1991) 77; Briant (2002) 379–80. The foundation charter of Darius’ palace at Susa (DSf) describes Lebanese timber brought to Susa, presumably down the Euphrates, Briant (2002) 382. Conon’s voyage to see Artaxerxes II included a passage down the Euphrates from Thapsacus to Babylon (“κἀκεῖθεν εἰς Θάψακον τῆς Συρίας πορευθεὶς ἀνὰ τὸν Εὐφράτην ποταμὸν ἔπλευσεν εἰς Βαβυλῶνα”); Diod. 14.81.4; Briant (2002) 382. Strabo (16.1.11) notes that Alexander had boats constructed in Syria and sent down the river to Babylonia for his Arabian expedition, but that was a one way trip necessitated by the lack of appropriate building materials in Babylonia, Meiggs (1982); Briant (2002) 382, Arr. Anab. 7.20.2-4. Strabo (16.3.3) also reports Aristobulus’ claim that the Gerrhans, spice merchants from southern Arabia, transport their goods to Babylonia on boats (σχεδιαις) then they sail up (ἀναπλεῖν) the Euphrates to Thapsacus from where they distribute their goods (briefly discussed by Young (2001) 94). Another story which Strabo (1.3.1) relates as an example of the absurdity of a predecessor (Damastes of Sigeum, a 5th century BCE geographer and historian, Brodersen, ‘Damastes’, BNP) tells of a water voyage from Cilicia to Susa can be interpreted to show a river voyage down the Euphrates, Briant (1991) 78–79; Briant (2002) 382–83. See also, p.277 for evidence that Palmyrene traders used riverine transport for the Euphrates leg of their desert routes. 1000 Muller (1855) Geographi Graeci Minores (GGM) 1.244-54; FGrH 781; Schoff (1989) an exact reprint of the 1914 edition; Oelsner, Joachim, ‘Isidorus (2)’ BNP . Chaumont (1984) 64; Gawlikowski (1988) 77–78. 1001 Pliny NH 6.141: “In this place [Charax], Dionysius was born, the author of a very recent description of the whole world, whom the Divine Augustus sent into the East to compose a complete account for his elder son [Gaius] who was about to go to Armenia to manage Parthian and Arabian matters. This did not escape my notice, nor did I forget that in introducing my work I said that each author seemed most careful in describing their own situation. Nevertheless in this part I prefer to follow Roman armies and King Juba, in the volumes dedicated to the same Gaius Caesar describing the same Arabian expedition.” 293/448 are different scholars from the same place working on similar projects. 1002 In fact, their projects may have been quite different: Isidore seems to have written a history of Parthia, while Pliny says that Dionysius wrote a terrarum orbis situs. These works could have been as divergent as the primarily historical and geographical works of Livy and Strabo respectively , or as convergent as Herodotus’ ethno-geographical history . One argument in favour of equating the two authors is that Pliny’s list of sources for Book 6 includes an Isidorus but no Dionysius. 1003 However, while Pliny mentions Dionysius, it is unclear if he used his work as a source. In NH 6.141, Pliny specifically notes that he prefers to follow (“in hac tamen parte... sequi placet”) Roman military reports and Juba’s history for his own description of Arabia. Most of the Parthian Stations focuses on the first part of the journey , between Zeugma and Seleucia on the Tigris, the Parthian capital of Babylonia. 1004 This section of Isidore’s description takes the form of a list of named cities with intervening distances listed; for example: διαβάντων τὸν Εὐφράτην κατὰ τὸ Ζεῦγμα πόλις ἐστὶν ᾽Απάμεια, εἶτα Δαίαρα κώμη· ἀπέχει δὲ ᾽Απαμειας και τοῦ Εὐφράτου ποταμοῦ σχοινους γ\. εἶτα Χάρακα Σιδου, ὑπὸ δὲ ῾Ελλήνων ᾽Ανθεμουσιας <καλουμενη>, πόλις, σχοῖνοι ε\. μεθ᾽ ἣν Κοραια ἡ εν Βατάνηι, ὀχύρωμα, σχοῖνοι γ\. 1005 In this way , Isidore includes many towns but in little detail. He provides relative locational information for the sites mentioned in terms of distances between points on a line (“εἶτα Χάρακα Σίδου... πόλις, σχοῖνοι 1002 Chaumont (1984) 64: “Il n'y a aucun inconvénient majeur à admettre l'existence de deux personnages bien distincts Dionysos et Isidore, citoyens de la même ville, adonnés l'un et l'autre à la même science.” Gawlikowski deems the coincidence too unlikely and accepts Müller’s correction, (1988) 78. 1003 Pliny NH 1, libro VI: “Ex Auctoribus: M. Agrippa. M. Varrone. Varrone Atacino. Cornelio Nepote. Hygino. L. V etere. Mela Pomponio. Domitio Corbulone. Licinio Muciano. Claudio Caesare. Arruntio. Seboso. Fabricio Tusco. T . Livio filio. Seneca. Nigidio. Externis: Iuba rege. Hecataeo. Hellanico. Damaste. Eudoxo. Dicaearcho. Baetone. Timosthene. Patrocle. Deomdamante. Clitarcho. Eratosthene. Alexandro Magno. Ephoro. Hipparcho. Panaetio. Callimacho. Artemidoro. Agathocle. Polybio. Timaeo Siculo. Alexandro polyhistore. Isidoro. Amometo. Metrodoro. Posidonio. Onesicrito. Nearcho. Megasthene. Diogneto. Aristocreonte. Bione. Dalione. Simonide minore. Basile. Xenophonte Lampsaceno.” 1004 For the middle-Euphrates section of Isidore’s route: Musil (1927) 227–32; Chaumont (1984); Gawlikowski (1988); Edwell (2008) 11–16. Young (2001) 21–23 considers Isidore and Strabo to be describing different routes. 1005 Isidore 1: “Crossing the Euphrates after Zeugma (τὸ Ζεῦγμα) is the city of Apameia, then Daiara, a village; this is three schoinoi from Apameia and the Euphrates River. Then Charax Sidou, called Anthemusia by the Greeks, a city , five schoinoi. After that Koraia in Bartana, a fortress, three schoinoi.” 294/448 ε\”). Moreover, he usually tells us the kind of place he is describing, be it a town, city, fortress, river, et cetera, and often gives one or two other details. Examples of the former from this passage are Apameia (πόλις... ᾽Απάμεια), Daiara (Δαίαρα κώμη) and Koraia (Κοραία... ὀχύρωμα). An example of the latter is the longer description of Charax Sidou (“εἶτα Χάρακα Σίδου, ὑπὸ δὲ ῾Ελλήνων ᾽Ανθεμουσιὰς <καλουμενη>, πόλις, σχοῖνοι ε\”) for which, as well as the relative locational information, Isidore gives the note that the Greeks call it Anthemusia. Whereas Strabo provided little topological detail in a passage filled with movement, Isidore describes the same route as a static catalogue of stopping points and seldom explicitly notes the presence or effect of movement on the landscape he describes. However, like Strabo, his work has a strong sense of directionality as the orderly linear procession of his narrative implies movement from Zeugma towards Babylonia and beyond. The extant books of Ammianus’ history of the Roman Empire contain many geographical sections, and still more were included in the lost books, notably his description of Mesopotamia itself. 1006 As well as dedicated geographic digressions, his accounts of Mesopotamia drew on his personal knowledge of the region. While it seems that he did not accompany Julian on that emperor’s attack down the Euphrates, Ammianus’ description of that campaign is nevertheless very informative about the communities along the route. 1007 Like Isidore, Ammianus names many settlements and locations on the banks of the Euphrates, often with accompanying details on the nature and history of the places. 1008 However, unlike Isidore, his descriptions are embedded in a narrative of movement. While Ammianus’ account seldom refers to commercial activities, it reports the continued existence of many cities in the fourth century CE that Isidore had reported in the first century BCE. These survivals attest to a continuity of movement along the route, and most likely to its commercial use during that time. When the Roman emperor Galerius recovered 1006 Sundwall (1996). 1007 Matthews (1989) 13. 1008 Amm. Marc. 23.3.1-24.6.1. 295/448 control of Nisibis from the Sasanid king Narses in the treaty of 298 and made that city the focus of inter- imperial trade, it was likely the legitimate inter-imperial commercial prospects of the Euphrates route that suffered. 1009 The detailed accounts of the route which survive in Isidore and Ammianus reveal the important centres along the Euphrates in the first century BCE and the fourth century CE. However, this is their most important contribution to the understanding of the Euphrates as a commercial route. Ammianus’ account of Julian’s campaign focuses on military matters. Isidore’s itinerary is often said to be describing a trade route, but the text itself gives no indication of mercantile traffic or concern. 1010 As Gary Young points out, there is actually little evidence for traders using the Euphrates route, at least compared to the evidence for desert trade through Palmyra and the maritime route from India through Egypt. 1011 Strabo’s passage on merchant activity along the Euphrates is the best unambiguous evidence for trade along the Euphrates. 1012 As the most direct path between the important imperial centres at Ctesiphon and Antioch, it is highly likely that merchants used this route to move between those cities, both as final destinations, and as transit points for shipping goods further into their respective empires. Ammianus mentions a large fair at Batnae that was probably attended by merchants who travelled up the Euphrates from Babylonia, west from Adiabene and Nisibis, and east from the rest of the Roman empire. 1013 W e have no reason to doubt that regular commercial traffic occurred along the Euphrates route at least until 298 CE, except in times of war and crisis. 1014 1009 Petrus Patricius, fr. 14 FGrH 4, p.189, quoted in translation by Dodgeon and Lieu ((1991) 133) The passage is also quoted in translation and analysed, with a map, at Dignas and Winter (2007) 122–30 See also the Paikuli Inscription erected by Narses which mentions the peace (but not the terms) and the supporters of the King, Skjaervø and Humbach (1983). These treaties are discussed in Chapter 2. 1010 Young (2001) 7. The best example of this kind of description of Isidore’s text is the title of the main recent edition: Parthian Stations by Isidore of Charax: An Account of the Overland Trade Route Between the Levant and India in the First Century B.C., Schoff (1989). 1011 Young (2001) 188. 1012 Strabo 16.1.27. 1013 Amm. Marc 14.3.3. 1014 Young (2001) 189. In practice, the upheavals of the third century probably meant trade was somewhat sporadic after the second century. 296/448 These three descriptions of the Euphrates route show how the same path of movement could be treated in different ways depending on the selections and purpose of the author. The lack of mercantile activity in the accounts of Isidore and Ammianus Marcellinus highlight the commercial focus of Strabo. Moreover, the interaction with the places along the route as described by the former two authors lies in sharp