It is convenient to consider a range of modifications to the secondary fort within this notional third phase, although it should be stressed that there is no guarantee of contemporaneity.
The main outer walls of the structure showed evidence of reconstruction at the south-east corner, where a different type of sandstone had been used, cut into slabs, rather than the blocks found elsewhere. Where the south wall crossed the line of the backfilled ditch, some slumping was evident in the replacement walling. The same stone had also been used to provide a new north-south partition (191) in the southern half of the building, although only the lowest course survived and the western face appeared to have been disturbed. This was 4.70m long and 0.80-1.00m wide and terminated at the point where it would have joined the east-west partition (140). Angular fragments of the same stone, perhaps debris from dressing, were found in the makeup (173) of the floor surface in the newly-created south-east room, and in the fill of a north-south slot (156) in the northern half of the building.
Within the north-east room of the previous phase, the drain and parallel slot had filled with sewage matter and bone, the east-west slope in the room causing ponding immediately inside the east wall. As part of the preparations for Phase 3, dressed and undressed stone in a clay/loam matrix (188) was dumped in this low-lying area as makeup for a new floor (192). This floor was made of re-used paving slabs and faced stone rubble (128) and it seems to have incorporated the lowest course of the now levelled north-south partition wall (145). The flooring material included a pivot stone (see Coulston below).
The fact that both the new north-south partition and the new flooring in the northern half respect the line of the east-west partition (140) strongly suggest that it continued in use. In the southern half of the building, the makeup containing the angular debris (171) lay beneath a mixed dark loam with stone inclusions (157, 172), which was in turn covered by a cobbled floor surface (143). These cobbles lay beneath the north-south partition (191) and butted against the east-west wall (140).
At least one of the postholes identified (260) belonged to a likely rebuild of the contubernium area of the barrack and floor levels and hearths above those of Phase 2 were noted. A fragmentary flagged floor (201) was associated with hearth 204, 0.50m in diameter, and a dark brown, ashy loam (228), which presumably derived from the use of the hearth. These probably belonged with Contubernium 1, but 3.90m to the south of 204 was 209 (measuring 0.70 by 0.44m), another hearth with its associated patch of ash.
The via vicinaria was given a new surface (236, 312, 331) similar to that of Phase 2, cobbled with large round boulders as kerbs, although these were now to the west of the line of the old street, suggesting a slight widening of the surface. The new surface was bedded in a shallow (0.03m) deposit of dark brown/black clay/loam (336).
The partial rebuilding of Phase 3 seems to be indicative of neglect of the fabric of the officer's quarters towards the end of Phase 2: the drain in the north-eastern corner having overflowed to leave an area of organic waste; the evident damage to the southern end of the stone building; and the presence of dressed stone of the type used in Phase 2 in contexts associated with Phase 3, such as the rubble raft placed over the drain deposit.
The interior arrangements of the building were changed, although the central east-west partition was retained. The removal of the Phase 2 north-south partition in the northern half led to a larger, flagged room being formed, a new partition being inserted instead in the southern half.