A watching brief was carried out by Colchester Archaeological Trust at 97 High Street (Castle House, Castle Bailey), Colchester, during the machine-excavation of nine test-pits (each c.10m x 0.6m in size) by the developer.<1> The test-pits varied in size, but were usually just over 1m in length and at least 0.6m wide. The remains of the south precinct wall of the Roman Temple of Claudius (MCC1544) were uncovered close to the modern ground-level in the southern part of the site. In the northern part of the site, the ground-level had been considerably made-up in recent times, and significant archaeological deposits were only reached in TP3 at a depth of approximately 1.75m below the modern ground-level. Part of a large Roman wall or foundation (F1) was uncovered in TP6. This was constructed of a hard pale brown mortar containing fragments of stone and Roman brick/tile. It lay approximately 0.4m below the modern yard surface. No edges were revealed to confirm its alignment. However, it probably formed part of the south precinct wall of the temple. Its proximity to the modern yard surface suggests that F1 may have belonged to a blocking wall or pier-base on top of the foundation platform rather than to the platform itself. Part of a hard, pale brown layer of mortar (F2) was uncovered in TP8. This contained stone fragments and lay 1.05-1.1m below the modern yard surface. It was not possible to examine F2 in detail, due to the proximity of asbestos. No edges were visible. The position and depth of F2 suggest that it probably formed the top of the foundation platform of the south precinct wall, close to its southern edge. The mortared surface F2 in TP8 was sealed by a brown gravelly deposit (L6), 0.30-0.35m thick. This contained a few fragments of stone, Roman brick/tile and opus signinum. It was probably Roman in date, but did not appear to be metalling. Similar layers were found during earlier investigations nearby.<2><3><4> They were perhaps make-up layers or material scraped up from the east-west street to the south. The gravelly deposit L6 in TP8 was sealed by post-Roman topsoil or ‘dark earth’ (L4). A brownish layer, similar to L6 in TP8, was also uncovered in TP7. It lay at about the same depth, 0.75m, below the modern yard surface. It was at least 0.30m thick and contained fragments of stone, Roman brick/tile, mortar and opus signinum. It was also sealed by post-Roman ‘dark earth’ (L4). Trowelling in the south-western corner of TP7 revealed some opus signinum at a depth of just over 1 m below the modern yard surface. This may have been in situ, and could have been part of an east-west Roman drain. Several fragments of Roman building materials, including part of a column brick and a piece of Purbeck marble, were recovered from the machine spoil from TP7. These were probably residual in post-Roman deposits. However, it is likely that the fragments originally derive from the south precinct wall or from associated Roman deposits nearby. The only significant archaeological deposit reached in the northern part of the site was in TP3, 1.75m below the modern ground-level. This consisted of a brownish layer (L5), which contained abundant oyster shells, some mortar fragments, and a few pieces of Roman brick/tile. It was sealed by ‘dark earth’ (L4). These deposits survived only in the northern part of TP3, as they were cut away over the southern part of the test-pit by the foundation trench for the existing east-west brick wall. The part of L5 uncovered in TP3 lay c.1m north of the Roman south precinct wall and 0.25m below the level of F1 in TP6. It was perhaps the fill of a medieval or later pit cut into the rear of the Norman rampart.