The east-west (140) partition wall was eventually demolished and the surviving, lowest course covered by the next floor surface, which seems to have extended over the whole interior of the building. The north-south partition (191) may also have been taken out of use and covered by it, but disturbance made it impossible to ascertain. The new floor was a yellow sandy clay (102, 116/153) with charcoal inclusions.
In the north-east corner, part of floor 192 was raised, and a length of the old east-west partition (140) rebuilt (with the second course set back from the original line), to form a feature 1.40 by 1.00m that, when excavated, contained a deposit of coal. This was associated with a dark brown layer with much burnt clay (130) and mixed ash, clay, and charcoal (136).
There was plentiful evidence for the demolition of the building. Tumble was found along the entire eastern side of the building (076, 254; see Fig.3, Sections 2, 3, and 6), parts of which (254) showed signs of quite deliberate toppling. Apart from dressed stone blocks, there was also an associated organic matrix which included amounts of animal bone and pottery. On the southern side, a line of friable tile fragments lay parallel to the south wall and 0.20m to the south of it. The constituency of this material was quite unlike other fragments of tile from the site, including the stamped piece (see below). Taken together with the association of iron nails with the tumble on this side of the building, this appears to be indicative of collapsed roofing material, although the amount of material was very small.
Two pits (059, 148) were dug in the via sagularis immediately outside the north wall of the building. 059 was located slightly to the west of the central axis of the building and was 1.00m square and 0.57m deep (Fig.3, Section 1). Outside the north-east corner of the building, pit 148 measured 2.00m north-south by 1.40m across; it was 0.46m deep over most of its bottom. Both contained animal bone, 059 with a greenish organic matrix (063) which was black (062) where it was permanently waterlogged. 059 also produced fragmentary pieces of decayed wood from its lowest fill (062).
A third pit (141) was dug against the inside face of the east wall of the building and this too contained rubble and animal bone.
The men's quarters of the old barrack were not retained and the only contemporary features noted in this area were an east west slab-lined drain (207) and a small area of cobbles. Although V-shaped and slab covered, the drain may originally have had straight sides which were subsequently distorted, but as found it was 0.16m wide and 0.20m deep. It was filled with a blackish, ashy loam (235) which did not appear to be the result of sedimentation, but rather deliberate dumping into the drain. The small, elongated oval patch of cobbles measured 2.00 by 0.70m and was not further investigated. Over the whole of the former contubernium area, there was a deposit of dark, 'garden'-type soil (041) which was rich in late Roman pottery, but contaminated by later reworking and so not securely stratified.
There was considerable disturbance of the road surface in the south- eastern quadrant of the site, some of which belonged in the later Roman period. A hearth (016) in this area produced an archaeomagnetic date of A.D.295-420 (see below). It consisted of a square of sandstone blocks, burnt red, and associated with a deposit of ash and coal (017) which lay directly on top of it. There were lenses of coal and ash (018) nearby and a narrow gulley (355), 2.20m long, 1.10m wide, and 0.17m deep, also showed signs of reddening. The gulley was filled with ashy silt (344) and coal fragments. Over this was a group of flags and smaller stones (339) which dipped from north to south into the middle of 355/344.
Adjacent to these features was another area of stones (338), which included worked stone; it was from here that the tile stamped NV was recovered.
Phase 4 saw the officer's quarters retained, although the area of the contubernia was, it seems, no longer covered by a structure, with just an east-west stone drain crossing it. Although certainty is impossible, given the amount of damage in the western half, the whole interior was apparently converted to one large room, with some sort of industrial activity requiring coal taking place on the eastern side.
The roadway was now also the scene of occupation, most of it incomprehensible in terms of the surviving archaeology.
Deliberate demolition may be indicated by the tumbled walls noted around the officer's quarters, although these could perhaps have resulted from natural collapse brought about by the deterioration of the rendering and subsequent decay of the clay-bonded walls. In this context, the pottery and animal bone associated with this destruction horizon may perhaps signal the onset of desolation.