common to the great majority, bones seemed to suffer little or no decay. Had they, on the other hand, been intended to be used for burials after cremation, it is beyond doubt that urns or other vessels deposited to hold the ashes would have been more frequent. As a matter of fact, only in a very small proportion were complete vessels discovered, and none of these appeared to have been employed as receptacles of the kind. The total number of pits or wells dug out was one hundred and seven, and to these may be added upwards of twenty others discovered, as Dr. Smith has recorded, in the railway cutting in 1846. Even so, the whole area was not exhausted, and it seems more than likely that pits remain untouched within the fort itself, and possibly in the annexes. In the south annexe the ground was closely trenched, and very few in that area can have escaped notice. The pits that were found were widely distributed. There were thirteen in the fort, three between the ditches to the south of the Retentura, seventy-nine in the south annexe, one in the baths, and eleven on the river bank to the north. As a rule, they showed no signs of having been laid down on a definite plan. They were sometimes set down singly, sometimes in groups.
There was much variation in size and in construction. The deepest were those close to the ramparts; on the south, Pit XVII had a depth of thirty-one feet nine inches, and on the north, Pit LXI went thirty-six feet down. None of those to the south of the railway was more than thirteen feet six inches deep. Comparatively few showed any indications of masonry. But fourteen, all of which lay in the south annexe, had remains of building, and had evidently served as wells. One of the wells (XCI) was found in remarkably perfect preservation. At the surface its diameter was nine feet. At a depth of three feet a stratum or floor of yellow clay was reached, having a thickness of eighteen inches. The clay itself rested on branches eight to ten inches in thickness laid over the well. In the middle of the floor was a small aperture, two feet square, the sides of which were lined with stones, which gave access to the well. Below the covering the well was carefully built, going down to a further depth of twelve feet. Another pit (LXXXI) had set into it a timbered framework, measuring three feet six inches by three feet, round the edge of the framework at the mouth was a thick layer of fine yellow clay, while lower down it was surrounded by stones through which water would percolate. A third (XLVII) had been sunk in soft ground. In this last case about a dozen wooden posts had been driven in vertically against the sides, while bunches of heather
107