Two trenches (T3 & T4), each c.3.4m long x 1.5-1.6m wide, were excavated in the northern part of the development site (to the north of the Roman arcade, in an area that was previously largely unexcavated), aligned N to S and c.8m apart. These were excavated to establish the depth and significance of the latest archaeological deposits in this area, and to retrieve dating evidence from them. T3 & T4 confirmed that the ground level in the northern part of the site had been raised by over a metre with modern make-up. The natural subsoil was not reached during the investigations in 2014.<1> An inhumation burial (F5) was partially uncovered at the northern end of T3 (most of it lay beyond the trench edge), most likely aligned E to W with the head to the west. Although there was no absolute dating evidence, the report suggests that the inhumation is likely to date to the 16th or 17th centuries, when burial of prisoners in the Castle bailey took place (see MCC9291). It should be noted, however, that at least one unfurnished inhumation burial (and possibly more) was defined below the Castle rampart on the west side of the temple precinct (see MCC2091 nad MCC2092). Approximately 0.5m to the east of burial F5, part of another feature (F6), probably a pit, was exposed at the northern end of T3. Like F5, F6 cut L7. A quantity of animal bone fragments was concentrated in a lens within the pit. Most of the bone remained in section, although a few pieces were recovered and these mainly came from a horse mandible; it seems likely that most of the bone in the section was also horse. No dating evidence was recovered from F6, but, like F5, it was probably post-medieval in date. Thick deposits of demolition and/or robbing debris survived in situ in T3 and T4. These deposits (L9 & L10) were excavated in T3, where they sealed the collapsed attached column F8 (see below). Datable finds from L10 were sparse, but there were one or two potsherds of probable 12th-century or later date. A small quantity of late 11th- to 12th century potsherds was recovered from L9 (which sealed L10). A thin, patchy yellowish-brown, clayey layer (L8), 80-100 mm thick, sealed L9. It contained fragments and flecks of chalk and patches of redeposited cover loam. It was perhaps upcast from the digging of the Norman ditch to the south. The eastern trench (T4) in the northern part of the site was not dug as deep as T3, and thick deposits of Norman demolition debris were not excavated. However the upper part of a layer (L12), which contained fragments of mortar and other Roman building materials, was exposed in the bottom of gully F7. This layer remained unexcavated, but it was probably comparable to demolition debris L9 in T3. A thin layer of yellowish sand, c.80mm thick, sealed L12. It was perhaps upcast from the digging of the Norman ditch to the south. In 2014, a narrow east-west trial section was dug across the middle of T4. This revealed a series of medieval deposits. On the eastern side of T4, a dark greyish, charcoal-rich layer (L3), approximately 150 mm thick, sealed the thin, sandy layer L11. The finds from L3 included potsherds of late 11th- to 12th-century date, as well as many fragments of animal bone and shell, mainly oyster shell. On the western side of T4, a shallow feature (F7) was excavated in the trial section. It was approximately 350mm deep and identified as a probable gully. A large quantity of finds was recovered from the narrow section across F7. This included potsherds of late 11th- to 12th-century date, many fragments of animal bone, including significant fish and bird bone assemblages, and oyster shells. The report suggests these deposits were perhaps located immediately inside the Norman inner bailey rampart. A large fragment of collapsed Roman (semi-circular) attached or half-engaged column (F8) was uncovered in T3, c.3.5m to the north of the arcade; it was probably from the upper part of one of the piers/columns found further south in 2015 (see ECC4220). The column had an estimated diameter of c.1m. Initially, the top of F8 was exposed in a small sondage that was dug in the bottom of T3. Subsequently, the western part of T3 was excavated deeper in order to uncover more of F8 and to investigate the deposits associated with it. The eastern side of the column was exposed for approximately 1.3m N to S, extending into the north and west sections of the trench. The fragment of column had a curving face and was constructed of alternate courses (at least nine were identified) of septaria and Roman column brick. At the northern end of the trench, the curved column fragment was found still attached to part of the face of the rectangular pier with which it had been built as one. Some fragments of plaster survived in places on the pier face and also at the angle where the column joined the pier. The limited dating evidence from the deposits of demolition debris associated with the column fragment F8 suggests that it fell in the late 11th or early 12th century. The western part of the same fragment of collapsed column was revealed in a small trench dug in 1964 (Trench III).<2> The remains in T3 and T4 were left in situ, given a protective covering and the trenches were backfilled. This work followed an evaluation (T1 & T2) in 2012 (ECC2988).<3>