Two 12m long x 1.5m wide trial-trenches were excavated by Colchester Archaeological Trust in June 2007 in an area currently used as garages next to Topfield, Rawstorn Road, Colchester (Phase II: a triangular-shaped parcel of land currently occupied by about 28 garages in a mix of concrete hard standing, rough ground and a tarmac access road). The investigation was carried out to evaluate the site of a known Roman inhumation cemetery; however, no burials were exposed by the evaluation trenches.<1> Roman features were present in the form of small pits of uncertain function in Trench 1 and one (quarry) large pit in Trench 2. Also, in Trench 1, there was an area of compact gravel that could possibly be interpreted as a metalled trackway or road of presumed Roman date. A quantity of Roman pottery (6,080 g) was recovered during the evaluation. Most of the closely-datable pottery is of 1st- to 3rd-century date. No Late Iron Age sherds were present. A few sherds can be dated to the mid 3rd-4th century (T2: L3, L10). The two sherds of shell-tempered ware (Fabric HD) from T2 (L10 and unstratified finds no 9) could date to the late 4th century. T2 featured a greater depth of post-Roman ‘dark earth’ than T1 (between 1.1m and 1.3m). The layers and features sealed by the ‘dark earth’ in T2 were considered to be upper layers of one large quarry-pit which had been used as a rubbish-pit later in the Roman period (F11). This pit took up the entire length of the trench and contained abundant Roman pottery, Roman ceramic building material including two pieces of roller-stamped flue tile, and oyster shell, slag, animal bone and lenses of charcoal. The edges of the pit were not seen within the evaluation trench, so therefore it must have been large. It was not bottomed but was at least 1.1m deep (2.4m below ground-level). Roller-stamped flue tile from this large pit hint at a connection with a public bath-house but not necessarily on the site. Slag within the pit suggests metal-working in the vicinity. A total of 1.669kg of faunal remains, consisting of 66 pieces, was recovered from the evaluation trenches. Remains were produced from ten contexts, mainly fills of a Roman date, although some material was recovered from fills of a modern date with the probability of being redeposited Roman material. T1 was initially machine-dug to between 950mm and 1m depth. 250mm of asphalt ca rpark surface and sub-base (L1 and L2) was removed to expose a 500- to 600mm thick layer of post-Roman dark earth (L3). At approximately 750mm below ground level, the dark earth gave way to a mid to dark greyish brown silty sand mottled with yellow sand (L4). The mottling may have come about through root action in the past. Flecks of charcoal and tile were seen in L4 as were larger fragments of Roman tile and pottery, plus some animal bone. The layer became thinner further north, giving the impression of hill-wash. This possible post-Roman hill-wash layer sealed a loose light yellowish brown sand mottled with dark grey silty sand (L5). L5 was quite prominent at the southern end of the trench and subsequent machine-digging revealed the layer to be approximately 250mm thick. However, at the northern end of the trench, L5 thinned out and became more orange in colour. Like L4, therefore, L5 could be interpreted as hill-wash. In T2, after the car-park surfaces (L7 and L8) had been removed, a post-Roman accumulation layer or ‘dark earth’ was encountered. This was a dark brown sandy silt containing occasional Roman tile, pottery and animal bone (L9), becoming slightly lighter further down (L10). These layers of ‘dark earth’ were thicker than that found in T1. At the eastern end of the trench, the dark earth was 1.1m thick but deepened to 1.45m thick at the western end of the trench. The trench was initially dug to a depth of 1.5m in the eastern end and 1.75m at the western end. At this level, a dark yellowish brown silty sand mottled with dark brown silty sand was encountered (L11). L11 contained much Roman pottery, brick and tile and some animal bone. The layer contained patches of slightly darker material without clear edges. Noticeable at the western end of the trench was a large spread of charcoal (F8) which was thought to be a pit cut into L11. On excavation of F8, the charcoal fill gave way to a yellowish brown silty material with a seam of oyster shell. Much Roman pottery, animal bone and some Roman brick and tile was retrieved from F8. No edges or base to the feature could be found. In the centre of the trench, a patch of oyster shell (F7) was noted. Before backfilling the trench, a scoop was taken out of the eastern end of the trench with the mechanical excavator to investigate L11. L11 was found to be 440mm thick and to seal a lens of charcoal (L14). L14 gave way to a sandy fill containing Roman pottery, oyster shell and tile (L13). At 2.4m below ground-level, natural ground had still not been reached. It was concluded that L11, L13, L14, F7 and F8 were not separate layers and features but were all fills of a large quarry-pit, occupying the whole length of the trench (F11). A watching brief was subsequently carried out during groundworks (CAT Report 480). <2>