The absence of immediately post-Roman evidence from fourth-century Roman military sites in western and northern Britain, forming a pattern of widespread disuse in the fifth and sixth centuries, has been strongly established. South of the Mersey the post-Roman finds from Segontium are not certainly of pre-eighth-century date, and the apparent evidence at Brecon Gaer has now been discounted by J.L. Davies. So – unless we accept a post-Roman dating for the penannular brooch from Castell Collen – only Pen Llystyn, where there is a single, possibly fourth-century, potsherd, has produced what might be considered convincing evidence of immediately post-Roman occupation at a fourth-century fort site. At Pen Llystyn the evidence, albeit enigmatic, seems to indicate fifth- or sixth-century reuse (perhaps even by the Irish notable named on a nearby Class-I inscribed stone) of the disused Roman fort for the site of a palisaded enclosure. It must, however, be doubted whether the site was in use in the fourth century. In the North, sites with fifth- and sixth-century evidence are only a little more plentiful – Manchester, Piercebridge, and Ribchester have possible, or probable, evidence of such use.