At the other end of the building the cold bath showed two distinct levels. The floor of the bath belonging to the lowest level has been mentioned above. It consisted of a bed of concrete, six inches thick, lying on broken red sandstone and river stones, with the step down and the moulding, already described, round the bottom. Over this floor lay a filling of clean dry river gravel, ten inches deep, and on this rested a second floor, three to four inches thick, of lime cement mixed with a little powdered brick, showing the same moulding round the edge as has been noted in the lower bath. The original floor of the frigidarium was about two feet above the lowest level of the bath in the apse, and even one foot above the raised level.

To the north of the frigidarium was another addition. Here an apsed room (J) had evidently stood; the floor of opus signinum remained lying on layers of debris with intervening lines of cobble stones, two feet nine inches in height. One of these lines of cobbles, twenty inches below the later floor, probably marked an earlier level: it was found projecting under the cobbles of the rampart eight inches beneath them. To the east was a wall running north and south, but returning at each end towards the west. This also seemed to be a late addition. It was founded upon two layers of river cobbles, bedded on seventeen inches of yellow clay, at which level, at the south-east corner six inches to the east, lay the rampart foundation.

Finds in the course of the Excavation

While the general destruction that has taken place renders it impossible to disentangle satisfactorily the different phases through which the bath buildings have passed in course of time, it seems clear that the enlargement and alteration of the original bath building had begun before the making of the rampart, and that further alteration took place afterwards. It is probable then, that we have here traces of the different phases which the baths underwent at four different periods, corresponding to the main alterations of the fort. The early building in all likelihood dated from the advance of Agricola. Its early extension and the large building on the west would belong to the first or second period of the later fort. The ditch and rampart would be constructed when the fort was reduced in area. Finally came the last occupation of all. Of the relics found in the baths, the pottery, though not large in quantity, was representative of the early, as well as of the later, period. The coins covered the same space of time as those found within the fort itself. The earliest was a consular denarius of C. Aburius Geminus (circa 129 B.C.), and