the third century. To the massive ruins of this fortress—ruins which tower over the marshes and break the horizon for miles round—we now turn.
   The ruins are fragments of the ramparts defending a rectangular fort. Three of the walls—south, west, and north—survive in singular completeness (P1. II). The fourth or eastern has vanished through erosion, although fallen pieces of it may still be seen at the foot of the cliff, and part of a curious foundation, to which further reference will be made, remains upon the brow. How far stable land extended seaward in Roman times we do not know. At Reculver, as we have seen, the fort was several hundred yards from the coast, and clearly depended rather upon the channel of the Wansum for approach by water. At Burgh Castle, on the other hand, the larger part of the fort occupied high ground, but the water-front was brought down to the water’s edge. Richborough was probably more like Reculver than Burgh; for the smallness of the surviving portion of the large legionary camp underlying the fortress is clear indication that much land has been lost here since the Roman period. Nevertheless, the Richborough fortress is 20 ft. below the summit of the hill on which it stands, and accessibility from sea and channel, and doubtless, therefore, from the Channel Fleet, must have been a determining factor.
   The area thus enclosed within the walls now measures about 490 ft. by 470 ft., and, if squared off to its maximum surviving width, covers about five acres. There is a remarkable piece of evidence that this was the approximate size originally intended by the builders. Mention has already been made of a foundation of which a part still remains near the brow of the cliff in the northern part of the fortress. This foundation lies north and south, is from 13 ft. to 14 ft. wide, and can only have been meant to carry a heavy defensive wall. An eastern rampart in this position would give the fortress a convincingly square plan, with its northern and (possibly) southern posterns in the centres of their respective walls. But there are difficulties. On such a design the easternmost ‘intermediate’ tower on the north wall would come inconveniently near the north-eastern corner, which may be supposed to have had a tower of its own. A more serious objection is that a deep pit sunk anciently through the broad foundation contained 140 coins of which none was apparently later than the latter part of the third century; the foundation was therefore already ignored at that time. These difficulties, combined with the complete absence of any superstructure on the foundation, compel the inference that the rampart which it suggests was never completed and that the plan of the fort was modified during process of construction by an enlargement towards the east.
   The construction of the three surviving walls is uniform. They are 10½ ft. to 11¼ ft. thick, and occasionally even more; their original height cannot now be determined, but must have exceeded 25 ft. Their material is a rubble-and-concrete core with facings of coursed masonry, much of which is re-used from earlier buildings,³²  and reveals here and there an attempt at ornamentation in the disposition of light and dark stones. Externally from about 8 ft. upwards are double levelling-courses of tiles, which do not run through the core. On the inner side of the wall the facing has mostly been
   ³² Built into the north wall externally is, amongst much other second-hand material, a much-weathered carving of a lion.