How Did the Base Work?

We now consider how the military base area operated, as a zone where a large number of people lived and worked on a routine basis. On one hand, to function it required the affordances of its internal communications, connections with the civil town, and access to roads, river, and lands beyond the walls; on the other, there was a need for surveillance and control of activities within the base, and of movements across its boundary. The most obvious part of the base boundary (Plate XXII) is the substantial mud brick wall ploughed across four blocks from the city defences just S of Tower 21, and blocking Wall, A, C, and D Sts, with a gate established at B St. How the S boundary was defined E of D St has always remained an issue. If it was necessary to build a wall at the W end, why was this not simply continued all the way to, e.g., the S end of the Citadel? Across blocks F7 and F5 it seems that the boundary of the military zone simply comprised party walls between military and civilian-occupied structures. The same was true within block B2, by the Citadel, although the boundary probably comprised building frontages along Lower Main St. On the plateau, as the camp wall may have been a subsequent local enhancement, except where the amphitheatre formed part of it, the boundary may generally have comprised the rear walls of military-held houses lining the S side of 8th St—probably all properties from the city wall to H St. The course of the boundary along the W side of the inner wadi is unknown, but the base is suggested, as along 8th St, to have incorporated at least all properties lining the S side of the Wadi Ascent Road, if not encompassing all blocks on the wadi slope—in which case the boundary here may rather have comprised property frontages on K St. The base area was split by site topography into two major zones, the flat plateau, and the N branch of the inner wadi around the Citadel. Each was further subdivided.

What and Where? Revised Overview of Base Extent

Archaeological evidence indicates that, during the final halfcentury of the life of the city, the area directly annexed by the military was significantly larger than the original excavators realized. In addition to concentrations of soldiers around the gates and defences, and at various places within the ‘civil’ town, the military came to control a single continuous swathe of the urban interior, comprising the entire N part of the walled area from the W defences to the river cliffs, and extending as far as the S end of the Citadel, plus the floor of the inner wadi right down to Lower Main St opposite the (by Durene standards) showy C3 bath, which it also apparently built. This area totals c.13.5 ha (c.33 acres)—a literal quarter of the intramural area which today covers c.52 ha (c.118 acres, measured from the CAD plan of the city by Dan Stewart; both city and base were slightly bigger in antiquity, before loss of the River Gate and parts of the Citadel). In its final form, the base included several distinct zones (Pl. XXIII). The NW part of the city had become a military enclosure, bounded on the E side by a continuous wall down the W side of G St, incorporating the street facades of the E3 bath and E4 house. On the S it was defined by the ‘camp wall’ from the city defences to D St; with no sign of a wall across blocks F5 or F7, the perimeter between D and F Sts is inferred. It must be presumed that, as to the W, the 8th-St-fronting properties of the two blocks were taken over, but that the party walls comprising the boundary with civil housing to the S was not further elaborated. These lines converged on the amphitheatre, which formed the corner of the enclosure. This perimeter of the NW enclosure involved physically blocking Wall, A, C, D, and 10th Sts. A major entrance was on 8th St, at G St between the amphitheatre and the E4 house.

The Wadi Zone Campus, Citadel, and C3 Bath

From the junction of H and 8th Sts, which gave access to the twin main axes of the military base zone on the plateau, H St led S to the bulk of the civil town and ultimately to the Palmyrene Gate, the steppe plateau W of the city, and the roads W to Palmyra and NW up the Euphrates to Syria. The fourth side of the crossroads followed a curving course SE, down into the inner wadi, then snaking through the irregularly laid-out old lower town to the now-lost River Gate, portal to the Euphrates and its plain. Of most immediate significance is that the Wadi Ascent Road also linked the plateau military zone with what can now be seen as another major area of military control, in the old Citadel, and on the adjacent wadi floor. The N part of the wadi floor is now known to have accommodated two military-built temples, the larger of which, the A1 ‘Temple of the Roman Archers’, was axial to the long wadi floor, which in the Roman period appears to have comprised one of the largest areas of open ground inside the city walls. This is interpreted as the campus, or military assembly and training ground, extension of which was commemorated in an inscription found in the temple. In 2011, what is virtually certainly a second military temple was found in the wadi close by the first, built against the foundation of the Citadel. This is here referred to as the Military Zeus Temple. Behind the Temple of the Roman Archers was a lane leading from the Wadi Ascent Road to the N gate of the Citadel. It helped define a further de facto enclosure, effectively surrounded by other military-controlled areas and so also presumed to have been in military hands. The Citadel itself, while in Roman times already ruinous on the river side due to cliff falls, still formed part of the defences. Moreover the massive shell of its Hellenistic walls now also appears to have been adapted to yet more military accommodation, some of it two storeys or higher.

Zooming In Rome, the Middle Euphrates, and Dura

Dura-Europos was a product and ultimately a victim of the interaction of Mediterranean- and Iranian-centred imperial powers in the Middle East which began with Alexander the Great’s conquest of the Achaemenid Persian empire in the later fourth century BC. Its nucleus was established as part of the military infrastructure and communications network of the Seleucid successor-state. It was expanding into a Greekstyle polis during the second century BC, as Seleucid control was being eroded from the east by expanding Arsacid Parthian power, and threatened from the west by the emergent imperial Roman republic. From the early first century BC, the Roman and Parthian empires formally established the Upper Euphrates as the boundary between their spheres of influence, and the last remnants of the Seleucid regime in Syria were soon eliminated. Crassus’ attempt to conquer Parthia ended in disaster at Carrhae in 53 BC, halting Roman ambitions to imitate Alexander for generations. The nominal boundary on the Upper Euphrates remained, although the political situation in the Middle East remained fluid. Rome long controlled the Levant largely indirectly, through client rulers of small states, only slowly establishing directly ruled provinces with Roman governors, a process mostly following establishment of the imperial regime around the turn of the millennia. However, some client states like Nabataea still existed in AD 100 (for overviews see Millar 1993; Ball 2000; Butcher 2003; Sartre 2005). The Middle Euphrates, in what is now eastern Syria, lay outside Roman control, although it is unclear to what extent Dura and its region—part of Mesopotamia, and Parapotamia on the west bank of the river—were effectively under Arsacid control before the later first century AD. For some decades, Armenia may have been the dominant regional power (Edwell 2013, 192–5; Kaizer 2017, 70). As the Roman empire increasingly crystallized into clearly defined, directly ruled provinces, the contrast with the very different Arsacid system became starker. The ‘Parthian empire’, the core of which comprised Iran and Mesopotamia with a western royal capital at Ctesiphon on the Tigris, was a much looser entity (Hauser 2012).

Developing a New Perspective on Dura’s Military Base

This research project arose, as many do, from an intersection of personal research interests and fieldwork opportunity. At its inception, I had already been working on material from Dura for twenty years, principally writing my PhD on the remarkable finds of (mostly Roman) arms and military equipment from the site, resulting in Final Report 7. I originally came to Dura as a Roman military archaeology specialist, but was acutely aware of my limited grounding in the specifics of the archaeology and history of the region. However, it is also clear that study of so huge and complex a data set as that from Dura must be a team effort involving many specialists from a wide array of disciplines and backgrounds, all of whom may bring outside perspectives potentially illuminating to the whole. My collaboration with MFSED began with an invitation from Pierre Leriche to examine some newly found items of military equipment. Spending time at Dura permitted an extended examination of the city, the Sasanian siege works, and Roman countermeasures (resulting in a publication on the Tower 19 complex, and indications of use of a ‘chemical weapon’ in the fighting: James 2011b), and especially of the military base where the soldiers whose equipment I had studied through artefacts and iconography had mostly lived. As previously mentioned, the base was not a primary research objective of MFSED. However, a project on the fixed infrastructure of the garrison would form a logical follow-on to my study of its martial material culture in FR 7. Contributing to MFSED’s general aims of recording and publishing the city’s remains, and to wider Dura scholarship, it also offered the chance to publish arguably the most important revealed but incompletely studied Roman military site in the empire. Further, this intra-urban military base constituted an ideal opportunity to pursue my own wider research interest, in how the Roman military interacted with civilian populations. At an early stage in my research career, I had come to believe that the Roman military could only be understood in context, of Roman society as a whole, and of the peoples it fought, conquered, and settled amongst.

Conclusion: Chiaroscuro

This project has explored the archaeology of Dura’s imperial Roman military base, and also considered other material traces of the presence of soldiers in the city, e.g. at the Palmyrene Gate and creation of urban baths. As such it here synthesizes the archaeological evidence of a literal quarter (or more) of this globally important archaeological site. It offers an example of the still under-appreciated potential of ‘legacy data sets’ and archival archaeology, and of resurveying ‘old sites’, to generate significant new knowledge, making best use of limited resources. It also considers ‘legacy ideas’ as well as more recent publications to generate new understandings of garrison, base, and city. I hope that this volume will further constitute a useful contribution to the study of the Roman armies, and the soldiers in their ranks. I also hope that it will establish that the military aspect is a vital part of the story of Dura itself, especially for the Roman era, and that the military base and the people who lived in it cannot be treated as literally and figuratively peripheral to Durene studies. The foregoing presents what has been a visually led project, and also one of space and of movement within it. It was conducted through a combination of examining the largely image-based archival records of the Yale/French Academy expedition and direct observation of the fabric of the city, especially of the remains exposed by the original excavations as they were between 2005 and 2010. It has also generated entirely new data expanding the picture through geophysical prospection of the unexcavated portions of the base area and vicinity. Physically moving around the topography of the former city and, where it was still partially upstanding, through some of its spaces, provided many key insights. Others derived from considering plans, aerial photographs, magnetometry plots, and recent satellite images. Not least, interpretations arose from generating the new drawings, largely plans, featured in this book.

When? New Outline of Development and Chronology

Such was the extent of the base in its final form, before the disruptions of the 250s. But how and when did it reach this extent and conformation? As we have seen, the small number of direct epigraphic dates coming from key structures in the military base concentrate c.209–16. While the Yale expedition were aware that there had been resident Roman auxiliaries from the later second century, and also identified (mostly erroneously) some components as belonging to the 220s–250s, they took this ‘epigraphic dating horizon’ as indicating that the military base, from camp wall to principia to baths, amphitheatre, and campus, including creation of most military accommodation, was overwhelmingly a rapid creation of the 210s. This notion of a sudden military transformation of the urban fabric at that time supposedly reflected a radical expansion of numbers of the Roman garrison—resulting in traumatic shock to the city. The concentration of dated inscriptions from the military base does seem to constitute a tight ‘epigraphic dating horizon’ c.209–16, or indeed c.209–12 if the amphitheatre was really an opportunistic coda. It certainly represents a major military building campaign. However, it has been misinterpreted, and its significance exaggerated, especially in taking it to mark effective creation of the base. Central to the ‘epigraphic horizon’, of course, is the dating of the principia to 211–12, with rebuilding of the Mithraeum around the same time. Other components of the programme may be implied by epigraphic information. Notably the detail of the undated inscription attesting building of the A1 Temple of the Roman Archers and expansion of the campus plausibly fits in the context of the 210s. However, the Yale project team pushed interpretation of the epigraphic evidence much too far in employing other texts to date military structures. While the ‘camp wall’ may well also have been built c.210, the epigraphic argument for this, comprising an inscription of debated reading not even found in proximity to the wall, is flimsy in the extreme. Similar misuse of epigraphic evidence is seen in the case of the three altars to Dolichenus found in X7.

Who Lived and Worked in the Base?

Information about the specific imperial military contingents resident in the city, and their composition, comes from formal inscriptions, dipinti, graffiti, and Dura’s famous papyri, including part of the archive of cohors XX Palmyrenorum. The case of Dura’s garrison illustrates the validity of Millar’s call for a general review of evidence and interpretations regarding Dura-Europos (Millar 1998, 474). While the inscriptions still remain to be definitively published, it is sixty years since Final Report 5.1 on Dura’s papyri appeared, during which there have been a further two generations of general scholarship on the Roman military. These have seen fundamental changes in understandings of the subject, while several publications on specific aspects of Dura’s Roman military presence are also yet to be integrated into any wider reconsideration of garrison and city. Notably, Kennedy’s work has substantially revised understandings of the chronology and development of one of the major garrison elements, cohors XX Palmyrenorum (Kennedy 1983; 1994), while Edwell has effectively demolished the long-established wisdom that the garrison was, in its later decades, under an officer called the dux ripae, supposedly a regional commander foreshadowing the territorial duces of the Dominate (Edwell 2008, 129–35). Dura’s military presence also needs to be reconsidered against the background of broader recent developments in Roman military studies. Key is growing awareness of the importance of the ‘extended military community’, encompassing both soldiers and the many dependants who, it is now clear, routinely accompanied them. We will return to this aspect later. A fundamental restudy of the textual evidence for Dura’s Roman garrison is, then, overdue and needs to be undertaken by those with proper epigraphic expertise, but in its absence an interim review here is a necessary companion to the archaeological research on the base. Despite major subsequent discoveries such as the Vindolanda tablets (Bowman and Thomas 1983; 1994; 2003), the textual record for the Roman garrison at Dura remains unsurpassed by any other site, in its combination of scale, diversity of media, and detail. Some 60 per cent of Fink’s Roman Military Records on Papyrus comprised Durene documents (Fink 1971).

Military Presence around and beyond the Base Area

The C3 bathhouse, in the heart of the old lower town, the grandest Roman bath in Dura, was not originally envisaged to form part of the present project as it lay well outside the area previously recognized as forming the Roman military base. However, it has long been thought that this bath, and the similar facility M7 at the other end of Main St near the Palmyrene Gate, were constructed by the soldiers. In the case of the C3 bathhouse it also became apparent post-2010 that the zone taken over by the military extended all the way into the adjacent block B2, while the S frontage of B4 opposite the bath on Lower Main St may well have been the edge of the unified military base. Consequently, had it been possible to conduct further fieldwork as projected for 2012, full re-survey of the C3 and (for completeness and comparison) M7 baths would have been undertaken. Circumstances prevented this; however, discussion of these facilities is still necessary, especially as both were only summarily published. This review draws on a preliminary inspection undertaken in 2010, as well as archival material. It concludes that the C3 facility was significantly larger and grander even than already understood. Like the other bathhouses, the C3 facility was visible as an upstanding mound, and especially noticeable as the Ottoman road ran directly through it. The bath was partially excavated during the sixth season, and archival records are relatively good including a six-page typescript document, ‘Notes on Roman Bath in Block C3, excavated Jan.–Mar. 1933’, annotated ‘M. Crosby?’ There is also partial photographic coverage of the structure. The finds registers record 262 finds from the building, over 100 of them ascribed to specific rooms; however, few can now be specifically identified. Brown published a brief account of the bath, composed from ‘the notes of the excavator, Miss Margaret Crosby, carefully checked by personal observation of the author’ (PR 6, 95, n. 7). This, the only publication on the C3 facility, formed part of Brown’s comparative study of Dura’s bathhouses (PR 6, 95–104 and pl. IV, reproducing archive drawing Bath N.12; note on the mosaics and inscription 631: PR 6, 104–5; pls XVI.2–3; XXXIX.1–3).

Project Context Rediscovery and Exploration

The ruined city known locally as Salhiyeh was virtually unknown to western scholarship until the twentieth century (Sarre and Herzfeld 1920, 386–95; Kaizer 2017, 64), but its ancient identity remained unknown until the aftermath of the World War I when collapse of the Ottoman empire saw Britain and France divide up much of the Middle East between them (Velud 1988; Barr 2011). As we saw, during operations against Arabs resisting the new western occupation, British-commanded Indian troops bivouacking at the site dug defensive positions and accidentally revealed wall paintings. These were seen and published by visiting American archaeologist James Henry Breasted (Breasted 1922; 1924), who first identified the ruins as those of the historically attested but unlocated ‘Dura . . . called Europos by the Greeks’ (Isidore of Charax, Parthian Stations, 1). The site thereafter fell inside the newly imposed borders of French-controlled Syria (Velud 1988). More substantial excavations were conducted and published with exemplary speed by Franz Cumont in 1922–3 (Cumont 1926), paving the way for the great Yale University/French Academy expedition overseen by Mikhail Rostovtzeff. This ran over ten seasons: (Dates from the Preliminary Reports, and Hopkins 1979, xxii–xxiv, except ninth and tenth seasons from information in Yale archives provided by Megan Doyon and Richard A. Grossmann.) With a Roman military presence attested from the outset, further traces were encountered throughout the city’s exploration, with the heart of the military base area being identified and excavated in the fifth season, and the great ‘Palace of the dux ripae’ in the ninth. While masterminded by Rostovtzeff, and more nominally Cumont, these giants actually only briefly visited the excavations on a couple of occasions. The dig was conducted under a series of field directors: Maurice Pillet, Clark Hopkins, and finally Frank Brown. These led a small team of American and European architects, artists, and archaeologists, mostly male (although women occupied prominent places on the team, including Yale graduate student Margaret Crosby and most notably Hopkins’s wife Susan); they were mostly young and inexperienced (including Hopkins and Brown).