ROMANO-BRITISH DERBYSHIRE nearly rectangular in shape, 8 feet long by 5 to 7 feet wide and 8 feet deep, walled with eleven courses of good masonry, floored with cement, and entered by eight steps (figs, i o, 1 1 .) The walling contained a fragment of an inscribed slab, dated about A.D. 158, which had been broken up and used as building material. The vault was therefore constructed or reconstructed some years later than that time. When excavated it was found to be full of mixed stuff. On the top were bits of building stone, fragments perhaps of the vaulting or the structures above it. Lower down were three other fragments of the inscribed slab, a drum of a column, a stone trough, a few corroded coins ascribed conjecturally to the fourth century, some potsherds, numerous bones, some of Roman date (bos longi- frons, etc.) and some judged by Professor Boyd Dawkins to be later, and at the bottom a wooden tub and broken concrete slabs, apparently the original flooring. The meaning of the vault is not doubtful. It corresponds to various vaults or sunk chambers which have been detected inside other forts in this island and abroad. These vaults lie under or very near the shrine of the headquarters building (p. 198), where the regimental standards and the military chest were kept. They have been reasonably explained as strong rooms, and in point of date appear to belong particularly to the late second and early third centuries. In its details size, shape, steps, position, and date the Brough pit agrees well with other specimens of these vaults, and we may fairly consider that it was built as a strong room. It may have been somewhat damp for its purpose, since in 1903 it filled quickly with water. But when the Roman drains of the fort were in order it was probably drier. Similarly the vault at Chesters, when first opened by the late Mr. John Clayton, filled at once, and a drain was required to clear it ; but it plainly was not below water level in Roman times. We may be tempted, however to think that ultimately, either in late Roman days or perhaps afterwards, it came to be used as a well, and the bucket at the bottom may be a trace of this stage in its history. Finally it became filled with rubbish and debris, and if Professor Dawkins has dated the bones rightly, a good part, if not all, of this process, must have occurred in post-Roman times. 1 Of other structures inside the fort or without it we know little. The excavators of 1903 revealed, but did not explore, a well-built edifice, almost touching the headquarters building on the south-east. Of other edifices in the fort nothing is yet known. Outside the fort we have, however, indications (as it seems) of the usual bath-house, placed to the south-east near Brough Mill and the union of the Noe and Bradwell Brook. Here Pegge saw in 1761 an oblong building with brick walls 1 Mr. Garstang, in his report, supposes that the pit was first built without steps ; then, long after A.D. 158, the steps were added, the centre deepened, and the pit converted into a well. I do not under- stand why the steps should not be here, as elsewhere, part of the original vault, or why they should have been constructed when the pit was converted into a well. The fact that the four lower steps are built up against the wall of the vault is not, in itself, conclusive of any view. The steps might at first have been wooden. Or it might have been desirable, for strength or dryness, to build the lower part of the walling round the vault continuously on all four sides. 205
