was a clay-lined well found at a distance of 20 ft. from the east side of the great concrete platform (described below). The contents of this well showed that it had been filled up within ten years of the Claudian invasion.
   How long the early military occupation of the site continued cannot yet be defined with certainty. It was probably short; it may not have extended later than about A.D. 50. It is at least clear that, by the last quarter of the century, the hilltop had been given over to other purposes. The earthen rampart which had presumably formed the inmost line of the camp defences was levelled and its materials thrown into the ditches. A road lying east and west was built, perhaps as early as the middle of the century, partly over the causeway of the camp entrance and partly over the ends of the ditches to the south of it (Pl. III). From this road another ran northwards in the direction of the northern postern of the subsequent fortress. In what is now the north-eastern corner of the fortress, was built a large dwelling—house, with walls of plastered flint and chalk, and with a small bath-system in its south-western corner. Coins of Vespasian, found beneath the floors of this house, together with late first-century pottery found in the refuse which subsequently covered them, suggest that the house was erected between about A.D. 8o and 110, and was destroyed not long afterwards.
   To the same period can now be ascribed the building of one of the most remarkable and at the same time one of the most puzzling structures ever found on a Roman site. It has been well known since Camden, in whose time it was called St. Austin’s or Augustine’s Cross; disputes have raged over it since the seventeenth century, and all the arts of the modern excavator have so far failed to find the complete solution of its riddle. But enough is now known to permit description and conjecture.
   The remains lie a little north of the approximate centre of the subsequent Saxon-Shore fortress. They consist of two parts, a substructure and a superstructure. The substructure is a solid platform, 147 ft. long, 105 ft. wide, and 5 ft. thick, composed of flint boulders and strong whitish mortar, and covered on top by a layer, 6 in. thick, of the same mortar which, at one point, was found in 1900 to retain a fragment of marble flooring. Four small holes, one at each corner, 4 in. square, ran through and even below the platform. Two of them, when found in 1843, contained (it is said) traces of wood; the other two, found in 1864, lacked this feature. Their use is difficult to understand, but they may have served for guide-posts in the construction. The platform is not merely a platform. It has a huge continuation downwards. In somewhat smaller size, 125 ft. by 82 ft., it has now been traced to its full depth, just over 30 ft. below the surface. This lower part consists of the same flint and mortar as the upper platform, and is clearly part and parcel of it. We may further call it solid, since eager excavators have bored 16 ft. into it on the east side without reaching any sign of an interior chamber.
   The superstructure is less simple. Its chief feature is a cruciform mass, about 5 ft. high, measuring 87 ft. by 7½ ft. along its north and south arm, and 46 ft. by 22 ft. along its east and west arm. The material of this mass is at variance somewhat with that of the substructure in the addition of ragstone, oolite and tufa, embedded in a brick-dust mortar which also differs from that below; it has also traces of a facing of small squared stones. Outside this ‘cross,’ on the same platform, are portions of a wall which