Internet Archaeol. 17. Willis. Approaches to the distribution of samian in and around buildings

10.2 Samian and intra-site analysis: approaches to the distribution of samian in and around buildings

Several publications have examined samian as a potential indicator of differing functional and status areas, differing activity zones and spaces. Studies have included work at the micro-level of specific buildings.

Birdlip Quarry
At the settlement site at Birdlip Quarry, by Ermin Street, Gloucestershire (Mudd et al. 1999) detailed attention was paid by the post-excavation team to the nature of the site finds as potential indicators of activities and depositional processes in and around the Roman period buildings. Samian was taken as one of the potentially key diagnostic indicators. Patterns indicative of usage and spatial organisation were searched for, and although the results were inconclusive, the main value perhaps of this work is as a methodological exemplar. A likely indicator of 'primary rubbish', for instance, was samian sherd size/average sherd weight, and it was noted that, 'In Period 1 the distribution of large sherds clusters around and inside the [circular] structures in Area A, with occasional sherds in the other occupation areas' (Mudd et al. 1999, 235). In terms of function, samian was considered a likely indicator of 'food consumption rather than preparation', and the distribution of individual sherds around the circular structure was accordingly examined; however, analysis showed that in terms of this proposition 'there is no obvious pattern to the distribution to support this', with large sherds of the Drag. 31R (?serving) bowl type, for instance, coming from midden and ditch deposits at the front of the building (Mudd et al. 1999, 247-9). It was concluded, in this case that, 'The distribution of individual categories of finds ... does not greatly aid the interpretation of activities carried out within the building possibly because there were no strong correlations between the types of vessels and the areas in which they were used, but perhaps more importantly because their place of final deposition does not necessarily coincide with their place of use' (Mudd et al. 1999, 249).
Shepton Mallet
Excavations at the roadside settlement at Fosse Lane, Shepton Mallet, in 1990, had involved wide-area investigation combined with one of the first examples of comprehensive spatial recording of finds at a site of the Roman era. The post-excavation team considered the distribution of artefacts in relation to structural remains, with samian identified as a potentially diagnostic category. It was concluded that, '... analysis of the plots for coins, samian and other wares showed that the former two tended to be associated with stone-built structures, rather than the timber structures indicated by other pottery clusters. While the samian and coin evidence could be used to suggest the differing status of the stone buildings, other pottery plots were less informative' (Leach with Evans 2001, 162-3). Examination of the finds per building identified some specific differences between the finds from individual rooms, as with the well-preserved stone-founded Building VII:

'Building VII in Period 2 [Antonine to early 3rd century] broadly reflected the pattern for the period as a whole ... However, Room B at least contained material which must have been residual by the end of this period so these differences may not be entirely functional. Room C produced ... a reasonable proportion of Central Gaulish samian ... 5% by weight ... The smaller Room B assemblage (167 sherds) was quite distinct. The room may have been used as a store. The assemblage contained no cooking pots, remarkably little BB1 (0.2% by weight), and no fine wares. It comprised mainly storage vessels ... Room A (341 sherds) produced an assemblage intermediate in character between Rooms B and C' (Leach with Evans 2001, 163).

Functional differences for these rooms were proposed on the basis of this information: kitchen (A), store room (B) and 'living room' (C). The finds from another stone-founded building (X), possibly an aisled hall with apsidal end, contrasted with the patterns at VII in Period 2. The function of the building, whilst potentially different, could not be closely characterised.

'The Building X group [Antonine] differed from that in Building VII, though derived from a much smaller excavated sample. A fairly narrow range of vessel types was present including beakers, tankards, cups, bowls, dishes and jars (36%). There was a high proportion of fine wares, in particular Antonine Central Gaulish samian ... and the assemblage also produced Dressel 20 amphorae (26% by weight). The bias towards tablewares rather than cooking vessels suggest a more specialised function' (Leach with Evans 2001, 163).
Site 1555, Snettisham
In the case of the rural settlement at Site 1555 on the Snettisham Bypass the spatial incidence of finds within the eavesdrip gully and other features associated with House 2 were plotted by feature segment (Flitcroft 2001, fig. 15). The occurrence of samian was specifically recorded. For whatever reason, samian sherds were shown to be particularly associated with the eavesdrip gully either side of the entrance to the building; this association with the front of the building was not a function of sample size.

Segontium

King and Millett's study of the incidence of samian from the 1975-9 excavations at the fort Segontium concluded that, in this case, the centurion's quarters did not show a higher proportion of samian than the contubernia, but they did have a higher proportion of decorated to plain ware (King and Millett 1993). This can be seen to support Breeze's observation from the evidence from the fort at Bearsden, that the centurion's quarters produce a higher quantity of high-status material when compared to other areas within forts (Breeze 1977). This presumably reflects the comparative wealth and status differences of these army personnel, and perhaps degrees of cultural affinity. Analysis of the Segontium assemblage also noted a high proportion of samian from the rampart back, which is 'not immediately comprehensible in relation to differences in status within the fort' (King and Millett 1993, 243), but which may relate to a location of meal preparation and consumption.

Location	Coarse Pottery	Glass	Samian	Decorated/Plain Samian	Aggregates
Contubernia	277 (87%)	9 (3%)	33 (10%)	2 dec/31 plain	319
Centurion's block	506 (84%)	39 (7%)	55 (9%)	11 dec/44 plain	600
Rampart back	38 (76%)	0	12 (24%)	3 dec/9 plain	50

Table 71: Summary of the incidence of coarse pottery, glass and samian (by sherd count) from selected locations within the fort at Segontium, 1975-9 (Source: King and Millett 1993, table 16.9)

Centurion's quarters
Hoffmann (1995) examined a sample of legionary centurion's quarters for evidence of the usage of space and organisation. Together with glass, jewellery and coarse pottery, samian was selected in this case as a potential index of 'different areas of use' (1995, 107). Hoffmann plotted the relative density of samian items and other finds per room (e.g. 1995, fig. 22) in order to isolate room function. The approach proved worthwhile. Work along similar lines in the future might consider the form/functional types of the samian and other vessels present.