Until the mid twentieth century, the study of Roman frontiers beyond continental Europe and Britain was largely the prerogative of military and colonial representatives. In Algeria, Colonel Baradez (Fossatum Africae), R.G. Goodchild in Libya, and most spectacularly, Père Poidebard (La trace de Rome), the Jesuit father who mapped, with the assistance of the biplanes of the French-Syrian mandate, a palimpsest of Roman to Umayyad sites across the Syrian pre-desert. Poidebard's map stops abruptly at the Turkish border and it has long been a desideratum of Roman frontier studies to map and study the traces of the Roman frontier on the Upper Euphrates. Timothy Bruce Mitford began his doctoral research on the Euphrates frontier initially under the supervision of Sir Ian Richmond in 1963. His fieldwork commenced at an opportune time since there were already plans to create a series of dams in the river's vertiginous gorges. Indeed, less than a decade later, the first of these at Keban flooded the only excavated fort on the entire line at Pağnik Öreni, which was never fully published. Numerous articles by Mitford have followed over the years, mostly concerning epigraphy and history, but the full exposition of his researches has had to wait. For much of his career he was a serving naval officer, but with great tenacity he continued an involvement in frontier research, often with a permit and representative from the Turkish ministry of culture.

The Cappadocian frontier was no ordinary Roman border. As with Syria, the Mediterranean superpower confronted Iran to the east, initially facing the Parthians and then, from the early third century AD, the Sasanians. In between was a patchwork of buffer states and client kingdoms, including Commagene with its great monument on Nemrut Dağ, and, most significantly, Armenia. Up to the mid first century, Roman control conforms to Luttwack's model of a ‘hegemonic empire’ to be succeeded by the direct territorial presence of legions and other garrisons at Samosata (Commagene) and farther north at Melitene (Eski Malatya) and Satala (Sadak) and later Trapezous (Trabzon). Roads linked the major garrisons, but the river often did not offer a suitable route as it forced a passage through steep gorges across the high ranges of the Taurus and Armenian Taurus. North from Erzincan, the frontier road struck off towards the Pontic Alps and the Black Sea. Few Roman frontiers presented such challenges of topography and extremes of climate. Yet the political significance of its neighbours ensured that we have extensive written accounts of Roman campaigns including the rare insight of the expedition of a provincial army under its governor Fl. Arrianus.

This large two-volume study, totalling 757 pages, presents a wealth of evidence and discussion. Beginning with chapters on the geography and history of the frontier, much of the first volume presents the evidence of Mitford's fieldwork, mostly on foot, supplemented by sightings from the air along the Taurus gorges in the 1960s and more recently over the Upper Zab from a military helicopter returning from northern Iraq. The course of the ‘frontier’ he describes is mostly defined by traces of roads and the remains of bridges, including the spectacular Cendere bridge, set back from the waters of the Ataturk Baraj and elsewhere long sections of the road agger still clearly preserved. The study begins at Samosata, now below 30m of water and never properly investigated before inundation. But the main contribution of this volume are the observations along the frontier track, the elucidation of the roads linking the major legionary fortresses, and the often separate defined track beside and above the river (better termed ripa than limes). Mitford's fieldwork took him to remote places with only a few accounts of nineteenth-century travellers as a guide. The predominantly Kurdish villagers mostly welcomed him in the 1960s, but the political conflicts over subsequent decades transformed traditions of local feuding into the bitter guerrilla campaigns between the Turkish state and the PKK (Partiya Karkerên Kurdistanê or Kurdistan Workers’ Party); this restricted access as well as creating official hostility on occasion. Unlike the experience of eighteenth-century visitors to Hadrian's Wall, journeying to inaccessible situations beside the Euphrates was not always rewarded by the discovery of extensive remains. Traces of mounds and small rectilinear structures on spurs are described at a number of places beside and above the river, assumed to be watchtowers and fortlets, but the remains of forts as indicated from written sources are often elusive. The course of the frontier roads is greatly enhanced by the detailed maps in Volume 2, including the clear presentation of annotated Turkish 1:200 000 maps dating from the 1950s. These are of enormous value in illustrating Mitford's narrative, but also for future studies of the frontier infrastructure.

The description of the frontier routes is no conventional limes guide, however. Early travellers’ accounts are extensively quoted in detailed footnotes, and woven into the main text is the parallel narrative of Mitford's own journeys, a portrayal of a now quite altered world. The illustrations reveal a similar eclectic range. Frequent images of the Roman army from Trajan's column illustrate (and celebrate) the Roman achievement; these are presented alongside images of Anatolian villagers and rural life (the best are in colour from 1963–1966). If the accounts of past travellers are constantly referenced, the same is not always true of more recent scholarship and certain omissions are significant. The opening and detailed geographic chapter includes a description of the Pontic routes across to the Black Sea, and here there is little reference to the major studies of Anthony Bryer and David Winfield, especially concerning the caravan routes to Trebizond. Similarly, in his discussion of the ripa north from Samosata, there is little reference to the preliminary survey with detailed maps by Mehmet Özdoğan (Reference Özdoğan1977), which includes descriptions of the aqueduct supplementing, if not at odds with, Mitford's account. In the same region Mitford's discussion of Tille and the possible site of the Roman fort differs from Stuart Blaylock's review of the results of David French's excavations, which explains that no fort was situated on the höyük, but traces were located in the fields to the south (Blaylock Reference Blaylock and Matthews1998). Despite these reservations, the author is to be applauded for not denying the region's recent history, notably by identifying those villages that witnessed the massacre and deportations of the Armenian population in 1915. There is also a valuable description of the region in late Ottoman times based on European accounts, a detailed documentation of the known Roman units, an inventory of inscriptions—including the carving of a centurion from Melitene on a rock close to the Caspian Sea—and a fragment of milestone from the flanks of Mount Ararat. It is to be regretted that none of the Greek and Latin texts are translated, as this would have made the book accessible to a wider readership.