A HISTORY OF NORFOLK east bank of that little stream and close to the water's edge. In shape it is roughly rectangular, longer from east to west than from north to south : its western side curves slightly outwards to the river (fig. 3). Round the enclosure stand, on the north and west, the striking ruins of ancient and massive walls (fig. 4), and on the east and south huge earthen mounds concealing walls beneath them, while outside, except along the river face, a great fosse or ditch can still be traced. The size of the place has been variously estimated, and indeed the earthen mounds hinder exact calculations, but the true internal measurements probably approximate to 1,400 feet from east to west and 1,100 from north to south, and the internal area seems to be about 34 acres.^ Four entrances can be observed, one in each side, but their antiquity needs to be confirmed by excavation. The walls consist of a concrete core, bonding-tiles and a flint facing, and in general resemble most of the Roman walls of towns and forts which still exist in our southern and south-eastern counties. They were strengthened externally by round projecting towers, such as were often used in the fourth century. One of these towers can still be traced close to the entrance on the west or river face : another, on the north face, was seen by Mr. Arderon, but is no longer visible above the surface. Both towers are said by the writer just named to have been constructed, like the walls, of concrete and bonding-tiles but to have been faced with bricks cut in squares instead of flints.^ In addition to the external towers, the walls appear to have had a ramp of earth behind them, but the exact relation between wall and ramp is not clear. Thus, it has been suggested that Caister was originally defended only by earthen ramparts and that walls were subsequently added in such a way that the outer face of the defences was new perpendicular masonry and the inner face the earlier earthen rampart. It may, perhaps, be simpler to suppose, in default of excavation, that here, as elsewhere, the ramp was erected for the sake of the wall.' The enclosure of 34 acres within these walls is now one wide enclosure of agricultural land, save that a medieval church stands near one corner, as at Porchester and Silchester. No recorded excavations have ever been made here, and we are totally ignorant as to the build- ings which presumably covered the ground in Roman times. We may be sure that the church does not represent a Romano-British sacellum^ as some rash writers have supposed, but we have no definite feature ' So the Ordnance Surveys and Wilkins in Archceologia, xii. 137. William Arderon, in Philosophical Transactions, xlvi. (No. 493) p. 200, gives the internal measurement as 1,176 by 792 feet, the internal area as 21 acres, and the total area including mounds and fosse as iz acres. Edward King (Munimenta Antiqua, ii. 49) and Gough {JJditions to Camden) follow him, but he is certainly wrong. I incline to think that he had figures for the internal area and for the space occupied by the mounds and fosse, and that he then, by some confusion, treated the internal area as if it included the mounds and fosse, and subtracted the figures for these latter when he should have added ; but the matter is not very clear nor, fortunately, very important.
- Mr. Fitch (see next note) says the facing is flint. He adds that the circumference of the western
tower was originally 31 feet.
- For descriptions of the site and walls, see W. Arderon (cited in the last note) ; Gough's Add. to
Camden, ii. 188 ; Wilkins, Archieolo^a, xii. 137 ; Gentleman's Magazine (1807), ii. 913 ; Edward King, Munimenta Antiqua, ii. 49 ; R. Fitch, Journal of the British Archceohgical Association, xiv. 1 24. The annexed plan is adapted from the Ordnance Survey. 290
