Internet Archaeol. 17. Willis. The incidence of samian ware in Roman Britain

13. Discussion and Conclusions

13.1 The incidence of samian ware in Roman Britain

13.1.1 Trends in samian supply, distribution, and consumption

Dated samian groups have been examined from a number of angles, and the information presented above by group. Synthetic analyses of these samian groups was undertaken, with examination by broad site category or type, a major strand of this project. It has been suggested recently (Burnham et al. 2001) that settlements can be defined primarily by the type of rubbish they generate and how it is deposited rather than by morphological typology. This is an interesting point. This study, however, has shown that there is a consistent and strong relationship between the nature of samian groups at sites and site type (as defined by morphology, function, location and scale). In other words, with regard to samian, the nature of deposited 'rubbish' is closely linked to site morphology. Overall, the incidence of samian across the province is highly structured, with site type a major factor. This is so despite the fact that each assemblage, from site to site, was the product of individual circumstances and choices; that there is structure and strong patterning in the distribution of samian indicates that within various communities there existed to a significant degree shared definitions, shared choices, like practices and options.

The differences identified by the project in patterns of samian consumption between contrasting site types, and presumably differing types of consumers, imply that both access to and perceptions of samian varied with site type/population. This aspect was highlighted too in the study of samian vessels present in burials (Section 9).

The Roman military were major users of samian, and consumed proportionally more decorated vessels, principally bowls (Sections 7.2 and 7.3). Evidently there were differences in the supply and distribution of imported samian between the military and civilian networks: disproportionately more decorated vessels were supplied to or acquired by the military. These differences are now well characterised by the data assembled in the present study.

'Small Towns', roadside settlements and smaller centres, together with rural sites, had been understudied categories within the Roman settlement hierarchy in the past (cf. Evans 1995a). An expansion, however, in work at these sites in recent years has been reflected in publications. These have generally included up-to-date methods in pottery reporting, with quantification, etc., such that our knowledge of the pottery consumed at these sites is much improved (as demonstrated in Section 7). This expansion in fieldwork and publication at these sites, in part due to PPG 16, provides a welcome correction to a previous imbalance in the available pottery data. It is likely that our knowledge of such sites will expand greatly in the coming years as much contemporary archaeological fieldwork at sites of the Roman era consists of either relatively small-scale trenching, or sometimes wider area stripping, and sampling of such settlement types.

The majority of the population, and of samian consumers, must have lived and worked at these sites. Overall analysis of the samian data shows that among pottery assemblages from both smaller civil centres and rural sites the proportions of samian occurring at both types of site are closely similar and low (cf. Tables 32 and 33). This differentiates these sites from major civil centres and military sites, which have much higher proportions of samian present (e.g. Tables 23, 24, 27 and 28). Likewise the ratio of decorated to plain forms at smaller centres and rural sites is reasonably consistent (e.g. Tables 35, 37 and 42). Rural sites have a slightly higher proportion of decorated vessels, possibly because more higher status settlements existed in rural locations (and are part of the sample analysed) as opposed to small towns which may have had a more utilitarian dynamic.

In aggregate the communities at these sites, that is both smaller civil centres and rural sites, must have been the main consumers of samian. This is so because despite the comparatively low proportions of this ware present in groups at such sites, the sheer number of these sites, which, when excavated, seem always to produce examples of samian, is so high. The investigation of Small Towns, roadside settlements and (non-villa) rural sites represents something of a key dynamic within Roman archaeology in Britain, and is beginning to change our understanding of the nature of the province.

A striking pattern was revealed by the project with regard to samian at smaller civil centres and rural sites. Pottery assemblages from such sites generally contain very low levels of samian, and within their samian assemblages proportions of ('expensive') decorated vessels are comparatively low; nonetheless burials associated with such sites as well as structured deposits feature samian prominently (Sections 9 and 12). With burials, the data indicate that it was more common for samian to appear as a grave inclusion in burials associated with rural sites and smaller civil centres than in cemeteries associated with major civil centres (as demonstrated and discussed in Section 9). Structured deposits at sites such as Alcester, Towcester and Heybridge, Elms Farm, feature samian, often disproportionately so, as opposed to other pottery types, while rural sites and features not infrequently yield rare, probably expensive, types (cf. 12). Hence a subtle picture emerges, in so far as consumption and turnover of samian at rural sites and smaller civil centres was less than at other types of site (cf. Tables Tables 23, 24, 27 and 28, 32 and 33), perhaps through careful curation; yet when the evidence is collated, as here, awareness of samian and conscious employment of it by these communities in culturally significant milieux is apparent. This suggests that samian was defined by rural communities and those living in smaller centres in a way that set it apart from other pottery. Low proportions of samian among assemblages from these sites hides a reality: it was 'valued' in these communities, just as it was treated differently by the military and populations in major civil centres.

Heybridge, Elms Farm, Essex, is perhaps typical in these respects. Samian forms very low percentages within pottery groups from this site, at around 1.5% by weight (cf. Table 30): it was not commonplace. This might be read as indicating a 'low status' settlement with little wealth available for samian and/or little interest in this Roman metropolitan tableware. Some samian vessels may have been in everyday use at the site. Others, perhaps the majority of vessels, may have been used only for special days and events. On the other hand, samian vessels are fairly well represented among the small number of burials associated with the site excavated to date (cf. Appendix 9.1). Samian is present in many instances of structured special deposits at the site (e.g. Well 6280; cf. Section 12), in some instances prominently. The evidence, in sum, indicates that samian was regarded differently from other ceramics at sites like Roman Heybridge. Samian was significant to these communities who exercised discrimination in its use, selecting it for special occasions.

Various lines of enquiry within the project have highlighted the nature of samian supply to the Roman army. Military sites evidently enjoyed regular supplies of samian and tend to have the most 'up-to-date' assemblages. For this, and other reasons, there will have been differences between the composition of a group of samian from a military site and that from a contemporary settlement such as a major town; this difference, of course, has implications for how we have dated groups from these sites. Among other indicators the distribution of Montans samian and of pre-c. AD 120 Lezoux ware in Britain (Section 6) suggest quite strongly that the supply of samian to the Roman military in Britain was organised on a regional basis, supporting Evans' military supply zones model (Evans in preparation).

The personnel of the Roman army were major consumers of decorated samian, mainly in the form of bowls. Military personnel will have been better able to afford decorated bowls, especially if they were officers. Decorated bowls may have been a 'required' accoutrement of the grammar of meals in forts and fortresses, especially among officers. If the majority of decorated bowls were owned by individual officers this would explain why they rarely show ownership markers, whereas plain wares, which may have been used more frequently by ordinary legionaries and auxiliary solders and which may have been stored, cleaned or transported together, show a high frequency of marking.

Various other aspects of the incidence and character of samian finds have been considered by the Projet such as wear and repair, with some clear patterns suggested by the data.

The database assembled for the Project allows other highly informative avenues to be looked at such as the 'life-history'/currency of specific types, levels of supply from various centres and the potential for dating samian groups via attention to group composition . The rich archaeological yield of this category of material is readily apparent from the analyses undertaken here. There remains though great further scope for studies at various levels with samian ware, not least at the level of inter-provincial comparison.