Internet Archaeol. 17. Willis. Lezoux samian ware cAD120+ in Britain

6.5 Lezoux samian ware (c. AD 120+) in Britain

6.5.1 The scale and importance of Lezoux samian (c. AD 120+) in Britain

Fieldwork has shown that virtually every site in Roman Britain occupied in the 2nd century AD produces sherds of Lezoux samian. The largest supplier of samian to Roman Britain was the Lezoux industry during its main export period of c. AD 120-200; the scale of this trade must have been huge (cf. Boon 1967, 30). Yet this is not apparent from Marsh's seminal paper which presents much quantitative data that suggests La Graufesenque supplied the greatest amounts of samian to London and to the province (1981, e.g. figs 11.5 and 11.7). Many published lists or site summaries of samian from excavations at military sites and major civil centres show greater amounts of La Graufesenque samian than 2nd century Lezoux ware. A case in point is Canterbury, where the samian evidence plotted over time by Bird for the Rosemary Lane and Gas Lane sites shows a marked peak for La Graufesenque samian, but a much lower frequency of Lezoux samian in the 2nd century (Bird 1982c, fig. 85). This profile is very similar to that obtained by Marsh in the case of London, and Bird observed that the 'outline of the graph follows a pattern which now seems to be fairly standard on Romano-British urban sites and may be as much a reflection of samian supply as of occupation of individual sites' (1982c, 158).

Phase 1 of this project had found sizeable samian groups of 2nd century date to be less common among published site reports than samian of the second half of the 1st century (Willis 1998a). Verulamium and one or two other sites were identified as exceptions, but generally the literature revealed that 2nd century groups were less frequent and tended to be smaller than those of 1st century date (cf. Willis 1998a, appendix A). This trend in part was a function of the considerable number of excavations undertaken examining Flavian forts/military horizons and the 1st century levels of towns which had been published, and an under-representation of smaller centres and rural sites. These publications, allied to the evidence assembled by Marsh (1981, especially 181), pointed to a lower level of importation of samian during the 2nd century than during the 1st. This has been a highly influential impression that was adopted in synthetic reviews of trade and exchange and in site discussions during the 1980s and into the 1990s (e.g. Fulford 1984, 124; Bird 1986a, 139-40). However, this impression is misleading. It was noted that if samian from small towns, roadside settlements and farmsteads, as well as from the northern frontier, is added to the equation the overall aggregate of Central Gaulish material would probably more than equate to that of 1st century South Gaulish ware (Willis 1998a).

Marsh had commendably assembled a large dataset, which was based on the information available at the end of the 1970s. It comprised samples from urban sites, with some military sites and Fishbourne (1981). His graphs generally show a consistent pattern of the frequency of samian at these sites over time and seem reliable in the case of the individual sites for which the graphs were generated. However, what was missing were samples from sites occupying the middle to lower echelons of the settlement hierarchy. By the late 1970s few of these sites had been investigated or published, reflecting an enduring imbalance in the types of Roman period sites examined in Britain that was to continue into the 1980s. As observed by some (e.g. Hingley 1989; Evans 1995a) this imbalance in the emphasis of excavations up until c. 1990 had implications for our understanding of the province. Marsh's sample reflected the nature of archaeology at the time (i.e. c. 1981). This situation has changed, with much more investigation of smaller centres and rural sites of the period, particularly as a consequence of PPG 16, and much new and dynamic insight into these types of site has been forthcoming. One consequence is that a much larger corpus of samian from such sites has been published. It is now clear that these numerous smaller centres and rural sites received and consumed markedly greater quantities of 2nd century Central Gaulish samian than 1st century South Gaulish samian. Reference to assemblages from small towns, roadside settlements and other smaller foci demonstrates this fact. This can be seen from the sites of this type in the database, as well as from a host of site reports appearing in recent years. The pattern can be seen, for instance, in the cases of Alcester, Warwickshire (Hartley et al. 1994; Ward 1996), Asthall, Oxfordshire (Mills 1997a), Heybridge, Elms Farm, Essex (Dickinson 1996a; forthcoming), Kenchester (Dickinson and Hartley 1985), Meole Brace, Shropshire (Dickinson 1994a), Neatham, Hampshire (Bird 1986b), Piercebridge, County Durham (Ward 1993), and Sleaford and Sapperton in Lincolnshire (Dickinson 1997d; in preparation). To take a specific case, at Fosse Lane, Shepton Mallet, Lezoux samian of 2nd century date accounts for 76% of all the samian from the site by count and 72% by number of vessels (Dickinson 2001b, 145). This pattern accords with the chronology of these sites as a class, since generally they had not taken shape until the early-mid 2nd century, from whence they evidently become vibrant centres (Burnham 1986; Hingley 1989; Millett 1990, 143-56; 1995). In addition, many rural sites, including villas, yield more (or indeed only) Central and East Gaulish samian than South Gaulish ware, largely, again, for very similar reasons. This is so, for example, at Hayes Farm, Clyst Honiton, Devon (Simpson et al. 1989), Rayne, Essex (Cheer 1989), Rudston, East Yorkshire (Pengelly 1980) and Whitton, South Glamorgan (Webster 1981). Moreover, Britannia as a province was effectively larger in the 2nd century than during the 1st century, with more potential consumers and consumer sites for Central Gaulish ware as opposed to the earlier South Gaulish samian. The main period of importation of Lezoux ware lasted for about 80 years, whereas that of La Graufesenque was a little less, at around 60 years. Hence, while it cannot be demonstrated unequivocally that there was more Lezoux samian imported into Britain in the 2nd century than there was La Graufesenque samian in the 1st century, it seems very likely that this was the case.

Further work needs to be undertaken vis-à-vis Roman military sites with occupation in both the 1st and 2nd centuries to ascertain whether amounts of samian from Lezoux generally outnumber quantities present from La Graufesenque. This is not straightforward, as a range of variables will determine amounts recovered archaeologically from these sites, such as depositional practices and garrisoning, both of which may not have been constant.

Mills, however, observes that, 'In general more Lezoux wares than Southern Gaulish products would be expected from a site occupied during the 1st and 2nd centuries AD' (Mills 2000, 48). Moreover, trends in the data imply that overall more Lezoux samian was consumed by the Roman military in the 2nd century than was La Graufesenque ware in the 1st (cf. Marsh's graph in the case of York (1981, fig. 11.8, no. 8)). To take a specific example, the histogram prepared by Dickinson on the incidence of decorated and stamped samian by date from Ribchester, Lancashire, shows, overall, significantly more 2nd century Lezoux ware to be present than 1st century South Gaulish ware (2000a, Table 16); Dickinson states that this is 'quite normal for a site of this type' (2000a, 204).

In sum, the matter of the overall supply of samian to Britain through the 1st and 2nd centuries warrants re-examination. Marsh's influential graphs must now be regarded as reflecting not the overall 'normal' pattern of samian supply to Britain and in north-west Europe (1981, 206), as he believed, but rather the pattern for the major towns and fortresses only. When the other types of site characteristic of the north-west provinces are considered the picture is very different.