The civilized world, the oikoumene, with the Mediterranean at its heart, co-existed with an outer world within which peoples, with distinct identities and different characteristics, were to be found. Some, like the Germans and Armenians, had been kept under scrutiny from the time of Augustus; others, like the Persians, from the sixth century b.c. nomads were, of course, harder to track and to assess, leaving room for stereotypes to arise between observer and observed. Hence the clichés and banalities which masquerade as descriptions of Huns and Arabs in the pages of Ammianus Marcellinus. Hence too a dearth of information about Berber tribes on the margins of Roman North Africa until the sixth century.

There was progress, though. By Late Antiquity, the Barbarian, originally a soft and indolent figure corrupted by luxury, later a hardy and irascible creature governed by emotion, had largely ceased to exist. With the loss of the west, the east Roman empire had to take stock of its neighbours and acquire accurate information about them if it was to supplement military action with effective diplomacy, properly targeted propaganda and appropriately flexible client management. On no sector of the frontier was this more important than that facing the Arabian desert, especially as two rival great powers, the Sasanian empire to the east, and the kingdom of Himyar in the south, could challenge and at times surpass Roman influence in Arabia. The scale of Diocletian's investment in frontier fortifications and communications testified to Roman awareness of the potential danger emanating from the desert, a danger which was to materialize on an almost eschatological scale at the end of antiquity, with the Muslim conquests.

Greg Fisher has therefore tackled a subject of undeniable historical importance. His prime concern is to trace the principal connections, political and religious, discernible in the sixth century between higher powers and Arab tribes. He focuses on three extended kindreds, Jafnids, Hujrids and Nasrids, which formed ruling dynasties in north-west, central and north-east Arabia. They were the designated agents of Romans, Himyarites and Sasanians, for managing Arab tribes within their respective spheres of influence (namely the Ghassan, the Maʿadd and the Lakhm). In a first chapter, he sets the scene with a survey of sources and a sketch of the geographical and historical setting. He is surely right to relegate early Islamic sources to the background given the many questions which can be raised about their value. The core of the book comprises three chapters dealing successively with the processes and consequences of Christianization (for Ghassan and Lakhm), of client management (for all three client groups), and of an increasingly clear pan-Arab identity (mediated by poetry, historical memories, language and script). Naturally, given the bias in coverage of the sources, he writes at greater length about the Jafnid dynasty patronized by the Romans and their relations with successive Roman regimes than about Nasrids and Persians or Hujrids and Himyarites. Making full use of insights into pre-Islamic Arabia garnered from Christian Robin and Michael Macdonald, he has produced a conveniently compact and stimulating study of pre-Islamic north Arabia, which, to a large extent, supersedes the multi-volume work of Irfan Shahid on the Ghassan. He concludes with a chapter on the deposition of the Jafnids who fell foul of Justin II and Maurice in the late sixth century and the abrupt replacement of the Nasrids as favoured clients by Khusro II at the beginning of the seventh, and with a final brief cast-forward to the Muslim conquests.

The picture presented of Arab tribes and ruling dynasties, principally the Ghassan and Jafnids, and of their relations with the nearby imperial powers, is developed out of careful examination first of key pieces of material evidence, inscriptions and buildings (in particular those with a documented connection to the Jafnids) and, second, of written accounts of important historical episodes. Fisher's main contention is that the Jafnids, like their Nasrid rivals, profited from great power patronage to build up their prestige and power on the margins of imperial territory, both among the Bedouin and among frontier provincials, and that high-level inter-empire rivalry could be exploited to increase their freedom of manoeuvre. It is irrefutable. So too his acceptance of the argument that the Nasrid and Jafnid courts played a leading role in the promotion of Arabic, Arabic literature and Arab pride in the sixth century.

There is, however, a somewhat abstract feel about Between Empires. The view is more that of a theorist of client management, acculturation and ethnicity, who is primarily interested in structures and processes, than of a historian preoccupied with particulars. So the account of events, which, as always, had a decisive influence on attitudes and policies, is impressionistic, given to illustrate rather than to help document the theses propounded. The four wars fought by Romans and Sasanians and the role of Ghassan and Lakhm in successive campaigns feature, but only in the background. The same is true of their diplomatic dealings. The context of international relations essential for explaining the withdrawal of imperial favour from both ruling dynasties is thus largely excised. The reader's understanding would also be much improved with a fuller description of the physical arenas within which Jafnids and Nasrids operated under imperial supervision, with particular attention paid to the geographical distribution of Roman and Sasanian bases. For it was a military presence in the desert margin, the badiya, and the burgeoning growth of townships with their own militias which gave Romans and Sasanians leverage on the ground.

Nonetheless Fisher has made a valuable contribution to the various historical debates he has joined, not least through his surveys, with bibliographical references, of the current state of scholarship.