actually planned, some traces of still older erections, while at Bar Hill a fort apparently of the Agricolan period lay beneath the fort of the Antonine age. Unfortunately at Camelon the different occupations have not been accurately defined. But both there and at Ardoch it was quite evident that there had been more than one.
So it had been also at Newstead, and, as the work of excavation progressed, it became possible to estimate the nature of the alterations which had from time to time been made on the original plan. From modifications of buildings, fillings of ditches, changes of levels, the story of the different phases through which the fort had passed was gradually unfolded. It will be convenient, even at the risk of some repetition of facts already stated, to bring together this evidence in the order of its discovery.
Evidence of two or more Occupations
Before the digging had proceeded very far, it became apparent that at least two occupations had to be dealt with. The foundations of the long buildings in the Retentura were clearly lying on disturbed soil. This gave the first indication, and the discovery of the ditch of the early fort passing beneath the clay rampart of the later occupations, and having at one point the drain of a later period built into it, provided ample confirmation. But it was not until the central buildings came to be investigated that alterations were observed which pointed to three and possibly four separate reconstructions, each of which suggested a distinct period of occupation. Finally the alterations on the earlier gateway of the later fort on the south supplied the evidence of probably a fifth period. The alterations in the buildings were most plainly marked in the Principia. The large hall which had been thrown over the Via Principalis was undoubtedly later than the outer courtyard to which it was attached. The walls of both were founded on cobbles, but at the point of junction on the north side, the cobbles of the west wall of the hall stood on a level with the top of the scarcement of the north wall of the Principia. Again, at the end of the inner courtyard the treasure vault in the Sacellum proved to have been no part of the original chamber. Its walls were built of old material, and were not in alignment with the walls of the Sacellum itself. Its foundations were lower than those of the room in which it stood, and in laying them the cobble foundations of the main walls had been exposed. The method of putting down these foundations by cutting a trench and filling it with river cobbles has already been explained. To remove the side of one of them was to weaken it, and it hardly seems possible that
78