Resumen:
The paper focuses on the problem of identification of Caer Urfe, one of the Twenty-Eight Cities of Britain listed by John Leland (d. 1552) from Henry of Huntingdon (d. 1155) after the ninth-century Historia Brittonum. Many of the twenty-eight have defied identification; but Leland¿s proposal of Tynemouth for Caer Urfe is now maintained by archaeologists on Tyneside, in the north of England. The author argues that Caer Urfe is to be associated with St Cynfarch, near Chepstow, in south-east Wales. It is one of ten Welsh religious communities named in the catalogue together with cathedral cities and ancient British hillforts, none of them on Tyneside. The paper also examines the case of Arbeia, recorded by Notitia Dignitatum as the name of the Roman fort at South Shields, Tyneside. The author shows that Arbeia has no link with Caer Urfe, nor does it mean `Arabs,¿ supposedly relating to the garrisoning there after 300 CE of troops from Iraq. Many Roman forts in Britain were called after streams close to them; Arbeia is hence best understood on the basis of Welsh erfin `turnips,¿ also the name of a stream near Aberystwyth, as `<fort by a> stream noted for wild turnips¿. The article is supplemented with an appendix containing a list of the twenty-eight cities from Leland¿s catalogue, with toponyms rectified after the twelfth-century Book of Llandaff and other Welsh documents
