There do, however, remain certain earthworks which seem to be laid out according to Roman rules, but which contain no traces of Roman habitations, are not connected with great Roman roads, and the banks and ditches of which are of greater height and depth than those generally in use among the Romans in Britain, and which therefore there seems reason to attribute to the post-Roman Britons. Such are Tamworth, Wareham, Wallingford, possibly Cardiff, though upon a Roman road, and the additions to the Roman works at York. The name Wallingford, “the ford of the Welsh,” may be quoted in support of this view. It is difficult to understand how it is that there are no remains in masonry which may be attributed to this period, for it is impossible that with the example of the Romans before their eyes, and a certain admixture of Roman blood in the veins of many of them, the Britons should not have possessed something of the art of construction. This difficulty does not occur in Gaul, whence the Romans were never formally withdrawn. On the Continent indeed, generally, buildings are found of all ages, from the Roman period downwards. Gregory of Tours, in his “Historia Francorum,” written towards the end of the sixth century, describes the fortified place of Merliar as of great extent and strength, in which there were included a sweet-water lake, gardens, and orchards; and M. de Caumont cites a description of an episcopal castle on the Moselle in the same century, which was defended by thirty towers, one of which contained a chapel, and was armed with a balista, and within the place were cultivated lands and a water-mill; and there were many such, like the defences of Carcassonne, of mixed Roman and post-Roman work, that is, of work executed before and about the fifth century.