By Eric Birley.
[Read on 24th February, 1932.]
Of the purpose of this structure there can, I think, be little doubt. At this point the fort wall is 6 feet thick; the overall dimensions of the structure are 12 feet wide by 7 feet from front to back, so that, in effect, it thickens the fort wall to 13 feet for a distance of 12 feet, giving a platform large enough for a heavy catapult to be mounted. The absence of bonding appears to support this interpretation: for the mass of clay and rubble, with a masonry revetment, would provide just such resiliency as was essential to take the recoil of such a catapult; had the revetment been bonded in, or the structure composed of solid masonry, the recoil might easily have shattered it.
Though there was no bonding, it was clear that fort-wall and angle-structure were contemporary; the masonry was similar, the mortar identical, and the mortar itself gave an answer to one of the previous year's problems; it was heavily charged with shale, and suggested that the shale in which the masonry at the north and east gates was found to have been laid, represents the same mortar, the lime of which had disintegrated as a result of its exposure by Anthony Hedley a century ago.
The rampart mound to the west of the gun-platform was examined with some care, and yielded interesting results; but the publication of them is held over for a later report.
Immediately to the east of the gun-platform was a building, of which the west end only has been uncovered as yet, similar to that found on site C in 1930 : close up to the fort wall, and built at a high level over the rampart-backing. The roughness of the masonry of its walls was surpassed by that of its flagged floor; it yielded coins of Trajan, Severus Alexander, the younger Tetricus (a barbarous imitation), and Magnentius, together with some indeterminate scraps of pottery; but there can be little doubt that, like the building on site C, it belongs to the last phase in the occupation of the fort.
It was found that the ditch (ditch A) was of the common military type, in which the sides become vertical at the bottom, leaving a straight-sided channel 1 foot deep by 2 feet wide. Further early pottery occurred in the bottom of this channel, including two more pieces of the samian bowl, no. 1 in the report for 1930. A few feet north of the point where the water-channel crossed ditch A, a second ditch (B), without the straight-sided channel at the bottom, was found to join the first from the southwest. The new ditch was traced for more than fifty feet, but heavy rain, that finally brought all work on this site to an end, prevented further investigation of it; it produced equally early pottery.
Inside ditch A (that is to say, to the east of it) there was a layer of puddled clay, some 12 feet wide, and inside that again a roadway of hard rammed gravel, some 3 feet below the present surface. The clay was scored by a number of sleeper tracks, and in the inner edge of it were several post-holes (marked by pegs in plate xxvii, fig. 2) with the points of posts still in them ; but in the small area opened up there was insufficient order discernible for the purpose of sleepers or post-holes to be deduced. The clay presumably marks the position of the rampart, round the corner of which ditch A is curving; and it might be expected that a wooden angle tower would occur at this point, with which the sleeper tracks might be connected. In one of them was found a large piece of Dr. 37 in the
style of MERCATO.
The layer of gravel was covered by a deposit, varying from 1 to 2 inches in thickness, of dark peaty matter; Dr. Blackburn was good enough to examine this material, and informed me that it appeared to come from the bottom of a pond. This was something of a puzzle, as the surface of the ground at present slopes uniformly down towards the south-east, and it seemed as though such a deposit could hardly have formed here. Early in June, however, a sharp thunder-shower converted the excavation into a pond, that was still full of water in October, when it was necessary to fill it in again. Clearly, at this point there was a depression in the original surface, where water accumulated after the early site was abandoned.
Before the water-channel was laid, the ditches had been filled in, and the whole area levelled up with material, presumably from the rampart of the early fort, if fort it was; that will explain why only the tips of the posts remained in the post-holes; the greater part of them had been shaved away, together with the bulk of the rampart itself. For the date of this levelling, there is no evidence as yet; the only finds of any note from above it were a denarius of Vespasian, in good condition, and a number of pieces of a Rheinzabern bowl, Dr. 37, in the later style of the potter IANVSpresumably dating from the time of Antoninus Pius at earliest.