In 1971 an Editorial in Britannia drew attention to a group of places in Roman Britain whose names imply fortification but where no pre-Romandefences are known and the town walls, if any, must have been built too late to provide an explanation. The most obvious example is Durobrivae (Water Newton). The identification here is not in doubt – it is established both by the Antonine Itinerary and epigraphically – nor is the meaning of the name: the first element, Duro-, means a place with walls and gates and the second element, -brivae (or -brivas), means ‘bridge’. But the town walls here can hardly have been added before the end of the second century and intensive examination of the area, on the ground and from the air, has produced no evidence of an important pre-Roman settlement, let alone a walled one, nor of a pre-Roman trackway such as might have required a bridge. There was, however, an early Roman fort here, and a Roman bridge over the river Nene, and the conclusion seems inescapable that the name, though Celtic, describes the Roman fort.