Since 1941 C.G. Starr’s monograph The Roman Imperial Navy 31 b.c.–a.d. 324 has been widely regarded as the standard English-language work on the topic. It terminates, however, with the battle of the Hellespont in 324 ce, in which Constantine defeated his last major rival, Licinius. Starr saw the battle as marking ‘the disappearance of the Augustan navy as an effective force’ (p. 198), with only ‘sparse evidence for naval activity in the Mediterranean in the fourth century’ (p. 208 n. 117). E.’s book sets out to challenge this perception and to demonstrate that Roman naval forces continued to be maintained in the Mediterranean for the next two centuries at least, if not longer.

After introducing his theme, E. begins by surveying the early history of the two Italian fleets established by Augustus and goes on to consider the direct evidence for their survival into the third century and beyond: the latter consists of a single dedicatory inscription at Misenum set up in 301/302 by a prefect of the praetorian fleet (CIL X 3344) and the listing of both fleets at their respective bases, but without their praetorian title, in the western portion of the Notitia Dignitatum (Not. Dig. Occ. XLII 7 and 11), which attained its final form around 425. E. suggests that the fleets were reorganised, reduced in size and deprived of their title following the defeat of Maxentius, whom they had supported against Constantine in 312, with a large proportion of the personnel being transferred elsewhere. The units are nevertheless assumed to have taken part in various naval campaigns recorded throughout the fourth century; contrary to E.’s assertions, however, the appearance of the Castra Misenatium and the Castra Ravennatium in the fourth-century Regionary catalogues of Rome does not in itself prove that those barracks still housed detachments of the fleets to operate the awnings in the Colosseum at that time. E. believes that the Misenum fleet probably disappeared sometime in the fifth century, but that the Ravenna fleet may have survived and become the basis of the dromonarii fleets of the Ostrogothic kingdom.

E. next turns to the fleets of the Western Mediterranean, the evidence for which comes almost entirely from the same part of the chapter of the Notitia, which records the continued existence of the Italian fleets (Not. Dig. Occ. XLII 2–21). He first focuses on two fleets that have no known predecessor under the Principate, one at Arles to the west of the Italian peninsula and the other at Aquileia to the east. These are interpreted as facilitating and protecting supplies and communications from the Mediterranean to the northern frontiers, from Arles via the Rhône and Saône towards the Rhine, and over the Julian Alps from Aquileia to the Sava at Emona and thence to the Danube. E. suggests that these and other fleets and units along the European rivers were established in the early fourth century by making transfers of personnel from the fleets at Misenum and Ravenna. The whole system will, however, have collapsed together with the northern frontiers in the first half of the fifth century. E. also hypothesises that, although the Notitia records no fleets in North Africa, boats would regularly have been attached to army units situated along rivers or along the coastline, and their crews could have been used to man the ad hoc campaign fleets that appear in our literary sources. He even goes so far as to suggest that the Vandals might have been able to make use of this infrastructure to build up their well-attested war-fleets when they moved into the area in the fifth century.

For the Western Mediterranean E. is failed even by the Notitia Dignitatum, since it lists no relevant fleets although they almost certainly existed. This omission means that we have no direct evidence for standing fleets in this region, apart from a single brief reference in the Theodosian Code, an imperial instruction to Auxonius, the Praetorian Prefect in the Orient in 368/369, ordering him to take responsibility for the classis Seleucena and all the other fleets of the eastern empire in order to bring their establishments up to strength (CTh 10.23.1). E. plausibly identifies the named fleet as a continuation of the classis Syriaca of the Principate, based at Seleucia Pieria, the port of Antioch-on-the-Orontes; the inclusion of this text in the Code does not, however, conclusively show that the fleet still existed in the fifth century, as E. suggests, since the instruction sets a precedent for the role of the Praetorian Prefect in the recruitment of all the eastern fleets. E. is nevertheless persuasive in his observation that after 324 there must have been a significant fleet based at Constantinople, and similarly that a continuation or successor of the classis Alexandrina is likely to have been maintained at Alexandria.

Finally, in Chapter 5, E. reviews the evidence for the types of ship employed in Mediterranean fleets, arguing that triremes and larger galleys had probably ceased to be used in any of the fleets by the late third century, which thereafter consisted of liburnae, which appear to have been two-level, 50-oared ships, or even smaller single-level vessels with around 30 oars, like the early second-century boats found at Oberstimmon on a tributary of the Danube, or those of the late-fourth/early-fifth century found at Mainz. E. argues that liburna thus came to be used in Latin for all types of warship until the word was replaced from the late fifth century in Byzantine fleets by the Greek dromon, although E. sees this as reflecting an evolution of types previously referred to as liburnae (or lemboi in Greek) rather than a major technological change.

Throughout the book E. carefully documents from the literary texts the occasions when fleets were collected and undertook operations, and draws sensible conclusions from these about the places where local standing fleets could have provided the core for campaign fleets, supplemented by ships attached to individual riverine and coastal units of the Roman army (although he overestimates the strength of the evidence for the latter even under the Principate), newly built ships and requisitioned civilian vessels. In the end, however, it is impossible to escape the problem that E.’s direct evidence for standing fleets is derived almost entirely from one part of a single chapter of the Notitia and a single reference in the Roman law codes. Even this could have been subjected to closer interrogation: the Notitia, for example, shows multiple instances of large units being split into smaller ones with the same name distributed over several sites; if the Misenum and Ravenna fleets provided the initial cores for those at Arles and Aquileia, why did their squadrons not retain their original names rather than take those of the Rhône (classis fluminis Rhodani) and of the Adriatic Veneti (classis Venetum)? E. suggests that units of barcarii could have taken part in coastal defence in North Africa, but barcae were shallow-draft lighters (Isid. Etym. 19.1.19), and all the units recorded in the Notitia were based on rivers or lakes, most of them far inland (Not. Dig. Occ. XI 22; XXXIV 42; XXXV 32; XL11 9, 15). Conversely, E.’s suggestion that the Rhône fleet could have been used to carry Count Theodosius’ army right across the Mediterranean from Arles to Africa in 373 gains credibility from the contemporary observation of Ammianus Marcellinus (15.11.18), overlooked by E., that, after it is joined by the Saône at Lyon and the Isère south of Vienne, the river can take ‘very large vessels, which are often subject to being tossed around by the winds’.

E.’s book demonstrates usefully and beyond doubt that the empire continued to be capable of conducting maritime operations in the Mediterranean as and when required throughout the late-Roman period, but also that the detail of how this was achieved remains essentially unknown.