Excavations by the Museum in 1934 (ECC815), ahead of the construction of a car park and new library, led to the discovery of a tessellated pavement situated in the north-west corner of Insula 35. Little else was found as the Roman levels lay 9ft beneath the surface.

In 1938-9 the groundworks for the Public Library (ECC234) were observed by Mr. E.J. Rudsdale when the NW corner of Insula 35 was found. Remains of a large Roman building were uncovered including an east-west wall which had 'enclosed a room with a tessellated floor '. East of the southern part of this floor were two foundation trenches indicating the position of two walls running N-S. Further east lay another patch of red tessellated pavement.<1>

During excavations at Culver Street between 1981-2 and 1984-5 (ECC337) the remains of a large Roman house were excavated. The building was situated in the north-west corner of Insula 35 adjacent to another building to the east (MCC922) and south of an east-west road (MCC733). This was an exceptionally large Roman house and is likely to have been one of high status as suggested by the generous proportions of its rooms, mosaic floors and evidence for expensive furniture in the form of a marble table leg. The building was almost certainly demolished at the end of its life, sometime between AD 275 and 325.

Excavation work on the building was limited to removing the surviving demolition deposits and examining the latest floors before stripping by machine. The building was roughly square and measured about 36 x 40m. It consisted of four ranges of large rooms around a central yard. The building contained at least four mosaic pavements, one hypocaust, glass windows, a kitchen and two cellars. Its high status was also indicated by the relatively large amounts of carved stone it incorporated including dados, wall veneers, cornices and a marble table. The majority of the walls appear to have been built of mortared septaria rubble.

The northern, western and southern ranges of the building each consisted of a row of rooms with a common passage on either side, the layout of the eastern side being unknown. A total of thirteen rooms and three passages were uncovered. Seven or eight of the rooms had sandy clay floors. One room had a floor of opus signinum, another contained a mosaic set within a border of plain red tessellation and one contained fragmentary remains of a hypocaust which had been extensively robbed in the early medieval period.

One of the rooms was a cellar with 0.65m thick masonry walls made of neatly-coursed septaria and tile. The west end of the cellar had been partitioned off with a masonry wall to form a small passage or room. Most of the walls of the cellar had been robbed out in medieval times. The cellar appeared to have been an insertion into the building and was entered via steps down from a passage with an opus signinum floor. The passage came from another room which appeared to be a kitchen. The kitchen contained at least eleven ovens and two hearths. A drain passed across the room at an angle. Other features in the room included two pits and some stake holes. Two of the other rooms examined had suffered badly from later cultivation as had one of the passages. One of the rooms contained a mosaic pavement largely destroyed in the medieval period. The other room and passage may have had sandy clay floors.

Another room was examined on a watching brief basis where very limited salvage work allowed the recovery of fragments of a smashed mosaic. The mosaic appeared to have been broken up in antiquity although there was no evidence to suggest it had been stripped. The room was sunken 2 metres below the level of adjacent rooms and may have fulfilled a religious function, perhaps as a shrine. One of the passages examined ran around at least three sides of a central yard. It had a tessellated floor. A second passage was situated on the north and west sides of the house and is likely to have formed public footways along the two streets (see MCC1076 and MCC1078) one of which continued along past the building to the east (MCC922).

A timber-drain lay around at least three sides of the central yard and passed through one of the passages. The drain appeared to have replaced an earlier one on the north side of the yard. Another drain lay diagonally across the yard and linked up with the other in the south-west corner and a fragment of a fourth drain was also recovered. The central yard had been lightly gravelled. Its northern half was destroyed in the late 3rd or early fourth century when a large pit was dug into it probably for sand and gravel extraction. Nearby was a wooden tank or box set in a clay lined pit and probably used for holding water.<1>