Group 7: the Severn Estuary

This group straddles the Severn estuary, naming sites in both England and Wales (terms which, of course, have no meaning in the geography of Roman Britain). Punctuobice was recognised by Richmond and Crawford (1949, 44) as a miswriting of *Portu Abone, Sea Mills, their most drastic emendation, while Venta S(i)lurum is Caerwent, both names occurring in the Antonine Itinerary Iter XIV, together with Cunetione, above 106₂₁.

The identifications of the three remaining names are more difficult. Metambala is possibly for *Nemetobala (Rivet & Smith 1979, 424, who wrongly credit it to Richmond & Crawford 1949, 41, who preferred *Nemetaballa) and it may perhaps be identified with the religious complex at Lydney, important until the beginning of the fifth century as the major cult-centre of Nodens, given the religious associations of the word *nemeton (‘sacred grove’). Dillemann’s (1979, 67) suggestion that it represents Μεταβολη (meaning something like ‘crossing’, a translation of the Traiectus of the Antonine Itinerary Iter XIV), while ingenious, cannot be right: there are no traces of a Greek source for the British section (Rivet & Smith 1979, 201 contra Richmond & Crawford 1949, 3 and Dillemann 1979, 64).

Iupania and Albinumno (for *Lupania and *Albiniano respectively (Richmond & Crawford 1949, 36; Rivet & Smith 1979, 247)) cannot be identified with certainty; whilst it is very tempting to identify the latter with the Ariconio, Weston-under-Penyard, of the Antonine Itinerary Iter XIII (and it is surprising that neither Dillemann nor Rivet and Smith were so tempted), the emendation required is not easy, and the form of the Itinerary is confirmed by the modern district name Archenfield (Welsh Ercing). Dillemann (1979, 67) connected it with Abone, Sea Mills, but this is more convincingly represented by the Cosmographer’s Punctuobice at 106₂₂, which Dillemann rather improbably saw as an error for the Pontibus, Staines, of the Antonine Itinerary Iter VII (Rivet & Smith 1979, 166). If we take the Cosmography to be more trustworthy than Dillemann would allow, *Lupania may have been one of the small towns at Rangeworthy or Nettleton, while *Albiniano may have been the site at White Walls. This latter identification raises the spectre of translation from a Celtic *albo- to English ‘white’, a phenomenon which may be less uncommon than has usually been supposed (Higham 1992, 202).