J Roman Stud 57

Contra Graham, a division in AD 197 was possible on different lines from the later ones, and did not necessarily foreshadow Severus campaigns of 209 onwards. Belief that his policy towards Britain was constant from 197-211 is unjustifiable. Conditions in Britain warranted repetition of the Syrian arrangement - a consular province incorporating the frontier, York and Chester, and a one-legion province lying back from the frontier and including Caerleon and London (the administrative capital holding the governor's residence). Parallels exist for an administrative capital distant from the main military force, and epigraphic evidence for the presence of speculatores further supports London's claim. Some capitals of praetorian provinces also had one legion as far from the capital as Caerleon from London. The northern military command would thus be politically less dangerous and the presence of consular governors there under Severus is explicable. As Caracalla withdrew from Scotland he may have treated Britannia Inferior as he then did Upper Pannonia (which presented no danger to Severus), reducing it to two legions simply by movement of the boundary to N of Chester. It is unnecessary to reject Herodian's statement or invent extraordinary command structures. PHWB

Review-article. The index and bibliography of this work are useful, but it lacks a map to illustrate the places discussed. The individual contributions provide stimulating accounts of the present state of knowledge. The reviewer discusses at some length the problems of the nature of the civitas, and also the vicus, in relation to the views of Reynolds, Frere and Mann expressed in this and other publications, and he comments on the status of Verulamium.

Review-article. This book is the first general study of origins and typology of Romano-Celtic temples since 1928, and its references to Continental sites are particularly useful in helping us to understand our own. Only free-standing temples are considered, household and military shrines and isolated altars being omitted. L points out the basic distinctions between Roman and native temple plans, but gives insufficient consideration to the close interplay of the two traditions, particularly in relation to differences between town and country practice. Data for 144 temples are presented in a chronological scheme; most of the plans are drawn to a common scale, though structural phases are not distinguished. For British sites, Type I of L's hypothetical reconstructions is preferable. His theories on the origins of Romano-Celtic temples are open to doubt in view of the very early date of the Heathrow structure which, pace L, cannot be ignored. It is arguable that the Roman contribution lay not in architectural form but in structural technique. The reviewer gives additional or corrected data for some twenty sites.