The training excavation run by the Dorchester Project team which commenced in 2008 in the allotments overlying the south-west quarter of the Roman town continued throughout July 2013. This was in the same 30 x 20m area as in previous years, with the main north-south Roman road through the town at its eastern margin. A pending change in allotment allocations, with the likelihood that part of the western end of the site will not be available for examination in 2014, meant that excavation was concentrated solely in the western half of the area with a view to elucidating as much of the overall site sequence as possible within this area. This involved the removal, in parts of the site, of substantial depths of deposit, up to 0.70m in places.

The natural ‘brickearth’ subsoil was revealed at a number of locations within the western part of the site. It was shown to have been cut by a variety of features, including probable pits and a number of gullies. A roughly east-west aligned pit (or possible ditch terminal) (3737) at the west end of the site contained a substantial finds assemblage including animal bone (dominated by sheep and pig) and, amongst the pottery, a large part of a fine white ware butt beaker repaired in antiquity. Further east a length of NNE-SSW aligned ditch or elongated pit (3724) had unknown relationships with two narrow east-west slots – although closely adjacent, none of these features appeared to intercut. Only short lengths of the east-west slots (3723 and 3733) were observed. South of these, however, a further, wider east-west gully (3746) appears to have been cut by feature 3724; again only a short length of 3746 was observed. A little further to the south again was another possible linear feature (3655), this time aligned roughly NW-SE, which contained a pottery assemblage dominated by local coarse wares and unlikely to date after c AD 70 at the latest. Finds from the other early features included a Nauheim derivative brooch and a strip brooch from gully 3733.

Towards the north-west corner of the site another possible early feature was a probable well, only partly examined. The upper parts of this feature were not clearly understood as they had probably been affected by sinkage before being largely removed by late Roman ditches. The main upper fill of the feature contained pottery of mid-late 1st century date. 

These earliest features were sealed (or, in the case of the putative well, probably sealed) beneath a substantial homogeneous deposit (3541/3574 etc) of brown clayey loam which was up to c 0.50m deep at the western margin of the site and extended at least 15m eastwards from that point; its eastern limit is not yet known. The finds from this layer have yet to be examined in detail so its date is unclear, as, equally, is its interpretation. One possibility is that it represents a dump or levelling up deposit, but the apparent homogeneity of the soil rather argues against this. Alternatively the layer may have formed as a gradual accumulation from the later 1st century onwards, suggesting relatively low-level use of the area at this time. If so, this might have interesting implications for understanding of the nature and intensity of activity in what was, after all, a fairly central location within the town. A further possibility, however, is that the layer formed as garden soil behind buildings fronting onto the main north-south road through the town, although evidence for such buildings is still elusive.

At the western end of the site, excavation of parts of a building already partly examined in previous seasons was completed. This structure was set upon the major build-up layer, probably not before the later 2nd century and possibly even later. Its construction was unusual; the superstructure, which must have been timber framed, was set mostly on irregular limestone boulders placed in a very poorly-defined construction trench cut into the build-up layer. The structure may have been initially in two parts – a northern unit measuring at least 8m north-south, and a smaller element about 2m north-south immediately to the south and with its ‘frontage’ on the same alignment (the exposed east-west ‘depth’ of both rooms being only 2m). The smaller unit contained a poorly-preserved hearth and was associated with substantial dumps of oyster shell immediately outside to the south, characteristics which suggest a possible interpretation as a kitchen structure, though some of the oyster deposits apparently predated the structure. A probable well (not yet completely excavated) immediately east of this structure was probably contemporary with it and would have been entirely consistent with this interpretation. In a subsequent structural phase the two rooms seem to have been linked to form a single building, the foundations being remodeled at this point. Deposits apparently contemporary with the use of the northern room produced coins of early-mid 4th century date and suggest continued use at this time. As previously suggested, the building seems to have been placed within a palisaded enclosure with an east-facing entrance roughly 3.5m wide broadly in line with the centre of the building facade, although the alignment of the southern part of the palisade trench/ditch was not very regular in relation to either to the building or the western edge of the road, from which the front of the building was set back about 24m.

Immediately east and partly in front of the enclosure was a small enigmatic feature, consisting of a sub-rectangular limestone and flint rubble spread with maximum dimensions of c 3m east-west and 2.5m north-south, partly set on a layer of large amphora sherds. The rubble was most clearly defined on its southern side, albeit with a slight but marked offset in this otherwise straight alignment, and had traces of northward extensions at its west and east ends, within which was a deposit of fragmented limestone roofing material. In addition, a single row of stones formed a small horn-like extension running south-westwards from the south-west corner of the rubble spread. A substantial part of a small grey ware beaker of later 2nd century type was associated with the feature and might possibly be relevant to its interpretation. The latter is unclear, however; while clearly significant, it is not certain if the feature carried a structure or, if so, what form this might have taken. The association with the entrance into the enclosure seems to have been deliberate, and stratigraphic associations indicate that the feature could have been contemporary with the primary phase of the building to the west.

The secondary structural phase of the western building was associated with the deposition of a substantial gravel surface. This was laid up to the frontage of the building but also extended out of the enclosure and was widespread across the northern part of the site. The gravel was laid over the rubble spread feature described above, sealing it, but it is notable that the southern edge of the gravel was laid only just south of the rubble base structure and never extended into the southern part of the excavated area; here there appears to have been relatively little complex stratigraphy until the commencement of deposition of the sequence of late Roman and later midden deposits examined in previous seasons.

North of the stone based feature the extensive gravel deposit overlay postholes and other possible structural features which may indicate the presence of a further structure in this area between the palisaded enclosure and the north-south Roman road, but further work is needed to clarify the nature of these features. Further deposits in the same area produced a number of interesting small finds including gaming counters and several brooches.   

A little further work was carried out on remaining parts of the late Roman ditches in the western half of the site. This confirmed that the upper parts of the probable 2nd century palisade trench had been recut in the later Roman period and had probably not completely infilled before the end of the period. As previously noted, the northern arm of the palisade trench was cut by two close set east-west aligned ditches with steep sided V-shaped profiles, the more northerly of which is only partly within the excavated area. It seems likely that these form a boundary (perhaps an enclosure) defining activity to the north. It is not clear if these ditches respected or removed the position of the north-east corner of the main building at the west end of the site. The building was certainly cut by another ditch, largely excavated in previous seasons, which ran parallel to the east-west ditches in the western half of the site before curving south-eastwards towards the line of the Roman road. As discussed previously, this ditch presumably defined an enclosure, part of which lay within the southern and south-western parts of the site.