but the evidence on which it is based might have formed a useful section of Mr. Randall's last chapter. The author on p. 95 speaks of a "quiet peasant migration" from Somerset and is inclined to date it to the Age of the Saints. In 650, however, Somerset was still in British hands and Anglo- Saxon colonisation of the area did not begin until at least 658. It would seem more reasonable to date any pre-conquest Anglo-Saxon settlement in the Vale to the two centuries preceding the actual conquest, when the hegemony of the Anglo-Saxon kings was being effectively exercised over the native kings of Morgannwg. On the strategic siting of the castles Mr. Randall is original and con- vincing but he has probably read too much into his distribution map of early castles. It seems unlikely that the conquerors of Glamorgan had any clear-cut strategic plan for the defence of the lordship during the initial stages of the conquest. Not everyone will agree, for example, that in the eastern part of the Vale "there was a complete absence of early castles, except possibly at Dinas Powis." Intensive farming in the more fertile Vale might well have eliminated the more primitive structures which have been allowed to remain in the less intensively cultivated "Border Vale" (cf. Merrick's eyewitness account of the moated house at Bonvilston, part of whose walls were burned "to lyme ground withall"). There is a strong presumption, too, that a number of surviving castles in the Vale whose fabric cannot be dated earlier than the thirteenth century cloak earthworks or foundations of a much earlier date. The siting of a castle was not always the result of purely military considerations as a recent authority on the subject has been at pains to point out (Brown, Medieval English Castles pp. 191-2). A castle certainly existed at Dinas Powis at the end of the twelfth century and was something more than a mere status symbol when the earl of Gloucester was massing troops for its reduction in 1222 (Cal. Pat. Rolls, p. 346). It is to be hoped that the present excavations at Llan- trithyd will shed more light on twelfth-century fortified sites in the Vale. Mr. Randall may be wrong, too, in the role he allots to Ewenny in the Anglo-Norman scheme of defence, if such a unified scheme existed. Ewen- ny priory has often been cited as a perfect example of a fortified monastery but as Mr. Ralegh Radford has shown, its fortifications are largely a sham, placed where they would most impress visitors, whereas "on the vulnerable east side the simple unfortified precinct wall remained unmodified throughout the Middle Ages." A brief resume of its chapters and an appendix of criticisms of subsidiary points scarcely does justice to the qualities of Mr. Randall's book. His discursive studies, the result of a long and careful study of both the written record and the landscape, are full of interesting observation and fresh discovery. His work has the further virtue of drawing attention to valuable work on Glamorgan which has appeared in journals not specifically devoted to local history. One such article may be singled out for special mention for its pervasive influence can be detected in much that Mr. Randall has written. Aileen Fox's article on the Dual Colonisation of East Glamorgan should be known by all local historians of Glamorgan whatever their special period. F. G. COWLEY