New identifications of Mediterranean imports continue to be made in Britain, permitting understanding the nature of this exchange to be revised. The recent publication of excavations at Bantham produced not only significant quantities of pottery (52 imported amphorae of diverse types and at least two LRC vessels of Hayes Form 3 (Bidwell et al. 2011), but an assemblage revealing some similarities to the emerging pattern seen on Atlantic sites. The amphorae include two examples of LRA4, as well as at least two amphorae of North African origin (Bidwell et al. 2011, 94). Perhaps more significantly, the assemblage is dominated by LRA1, with only a single body sherd of LRA2. The authors note that this might indicate a generally early date for the assemblage (Bidwell et al, 2011, 94; 112), while for Reynolds it indicates LRA2 and LRC did not necessarily travel together (Reynolds 2010, 110). The high proportion of LRA2 has been seen to mark the British assemblage as very distinct in character from that of the west Mediterranean, where LRA1 is typically the most common eastern type (Reynolds 2010, 106). The quantity of LRA2 in Britain and Ireland is certainly, relatively high, but LRA1 would appear to be the most common. Like Bantham, the site at Mothecombe produced a higher proportion of LRA1 (five vessels) to LRA2 (two vessels) (Duggan 2012). This emerging pattern might represent regional differences in supply or variations in the chronology of importation but might also suggest the British assemblage to be less 'unique' than previously considered, and instead, more closely aligned to Atlantic assemblages.
The 2011 Bantham report also proposed an earlier date than AD 475 for the first Mediterranean imports, possibly AD 450, and as a result raised the possibility of continuous, uninterrupted importation of commodities via Atlantic sea-routes from the later 4th century and throughout the 5th century (Reed et al. 2011 , 113). The authors note that North African and possibly 'Palestinian' imported amphorae have been identified in late Roman contexts at Exeter and therefore suggest that the Atlantic sea-routes were still open in the very late Roman period (Bidwell et al. 2011, 113-14). The presence of céramique à l'éponge at Exeter is also seen to reveal late Roman contacts with western France (Bidwell et al 2011, 114). They propose a model whereby the sites negotiating the exchange of minerals shifted between the late 4th and late 5th century, but although the sites receiving imported Mediterranean pottery changed, the routes of supply did not (Bidwell et al. 2011, 115). The narrow date range of c.AD 475-550 for the importation of Mediterranean pottery has also been questioned at Tintagel, where imports may again have arrived by the mid-5th century, and may, potentially, have continued beyond AD 550 (Barrowman et al. 2007, 332).
Another example from Britain with potential implications for the chronology, and indeed distribution, of Mediterranean imports is the recently reported identification of ARS at Pevensey Castle in East Sussex (Fulford and Rippon 2011). Although African Red Slip ware is an occasional find in Roman contexts up to the 4th century (Bird 1977; Tyers 1996, 152), its 5th/6th-century distribution has been seen as completely separate and associated with the new system supplying imported goods to western parts of Britain. This discovery, therefore, represents the site furthest east with late forms of this pottery, well beyond its established, post-Roman distribution. A body sherd of a possible East Mediterranean amphora may be associated with these finds (Fulford and Rippon 2011, 125).
One sherd of ARS was tentatively identified as the flange from a bowl of Hayes Form 91; variant C of this form has previously been recognised among western British imports at Tintagel and Dinas Powys (Thomas 1981, 8; Campbell 2007, 17), as well as at Bordeaux (Bonifay 2012, 253). Timby suggests an early to mid-5th century date for the Pevensey sherd, in line with earlier variants, 91A and B (Timby 2011, 145), whereas in western British contexts 91C is usually seen as a 6th-century find. A second sherd was matched to Hayes Form 99, which again has been identified previously, although variant 99C, as identified here, is potentially very late – indeed later than any other British or Irish examples of ARS. Thomas catalogued variants 99A and 99B of this form, including examples from Tintagel (Thomas 1981, 8-9). Within the Pevensey report a late 6th to 7th-century date is given for 99C (Timby 2011; Bonifay 2004, 179), although LRFW1 suggested production up to the later 7th century (Cau et al. 2011, 5). Unfortunately these two sherds are not illustrated, preventing comparison with other British finds. Two sherds were from the same vessel, the rim of which was illustrated in the report. It is described as being closest to Hayes Form 75, which has not previously been identified at any British site (Timby 2011, 145). The early to mid-5th century date given for this sherd is noted to be 'late amongst the British finds' (Timby 2011, 145).
Referring to Bird's study of African Red Slip in Roman Britain (Bird 1977 , 272), Timby suggests that these vessels are unlikely to represent 'traded cargoes' directed to the site, but instead might represent personal belongings (Timby 2011, 145). A similar discovery in western Britain would, doubtlessly, be automatically tied to post-Roman, long-distance import systems. The discussion chapter within the report, however, does describe these finds as altering the view that post-Roman Mediterranean imports are only to be found in western Britain (Fulford and Rippon 2011, 125). The authors also suggest a possible association with the previous discovery of DSP at the site (Lyne 2009, 101; Fulford and Rippon 2011, 125). It may be that the ARS from Pevensey raises the possibility of future identifications of Mediterranean pottery in post-Roman Britain beyond the traditional, western distribution. These sherds, however, demonstrate both the difficulties and importance in identifying known and datable forms based on incomplete or fragmentary vessels. As the chronology for the 5th- and 6th-century importation of Mediterranean pottery to Britain is largely founded on matching abraded fragments of fineware to published typologies, such attributions can have far-reaching implications.
The distribution pattern of the imported pottery has also been extended by the discovery at Rhynie in eastern Scotland of a small group of amphora sherds of types LRA1 and LRA2 (Noble et al. 2013, 1142). Excavations at this Pictish site also produced fragments of glass vessels imported from western France (Noble et al. 2013, 1142).
Overall, these recent publications allow the imported material in Britain to be better aligned with patterns in the west Mediterranean, and reveal that imported pottery in post-Roman Britain is both more varied and more widely distributed than traditionally assumed. It is clear that the later 5th and 6th century witnessed the unprecedented supply of east Mediterranean imports – including new types of amphorae and fineware – to a new group of sites in western Britain and Ireland. There remains, however, a level of uncertainty regarding the first half of the 5th century, and the potential continuation of late Roman patterns. As mentioned, LRA1 and LRA2 are not thought to be imported to Roman Britain, unlike North African amphorae and the East Mediterranean LRA3 and LRA4. Campbell describes the increasing identification of North African and Palestinian amphorae at late Roman urban contexts, including examples from London, Gloucester and Exeter (Campbell 2007, 19-22, 125-6). However, as these could not be confirmed as post-Roman imports he did not include them in his distribution. Typically, the North African amphorae found at these urban sites are of 3rd to late 4th/early 5th century types, and cannot be easily equated with the later African amphora imports identified in the Atlantic and west Mediterranean. The continuing use of the 'Bv' category has somewhat complicated this distinction.
Similarly, Bird's review did not record any forms of ARS that were necessarily 5th-century imports. The latest identified form, a base of ARS 67 from Southwark, was noted to be from a late 4th-century context (Bird 1977, 275). More recently, sherds of ARS were identified at Shadwell in London, but the identified form – Hayes 50/50A – is of 3rd/4th century date (Douglas et al. 2011, 177-9). This site also produced a number of North African amphorae of 3rd/4th century date as well as a spatheion type 1; the latter was found in a probable 5th-century context and might feasibly have arrived in the first half of the 5th century (Douglas et al. 2011, 68, 172; Williams 2011, 80).Three small bodysherds of LRA3 were also found at Shadwell, but it was not clear if these belonged to the earlier one-handled type or the two-handled type that characterises the post-Roman imports (Williams 2011, 81). Elsewhere, the presence of 'Palestinian' amphorae has suggested a general background of East Mediterranean amphora importation to Britain in the first few decades of the 5th century. An amphora recovered at Billingsgate in London, for example, has been considered to date to the first-half of the 5th century (Marsden 1980, 80-1; Campbell 2007, 125).
In the light of new Atlantic data, such as the amphora group from Bordeaux, as well as the recent evidence from Bantham and Pevensey, future considerations of the Mediterranean amphorae and fineware interpreted as late Roman imports to Britain have the potential to increase both the chronological range of the post-Roman imports and the extent of their distribution. Certain factors suggest, however, that the western British 'post-Roman' imports represent a separate dynamic – and that there was some break in supply via the Atlantic channels. Firstly, the shift in the focus of Mediterranean imports from urban sites to fortified, hill-top centres and coastal 'beachmarket' sites. Secondly, the lack of locations in Britain with Mediterranean imports of both late 4th/early 5th century date and later 5th/6th century types (LRA1, LRA2; LRC; late forms of ARS) (Campbell 2007, 126). An unprovenanced LRA1 was noted by Roberta Tomber from the Museum of London collections, but this was discounted as a 'genuine London find' (Tomber 2003, 107). The ARS from Pevensey also presents a possible exception, although it is feasible that these vessels arrived via an alternative, Rhineland, route (Fulford and Rippon 2011, 125). Finally, the evidence emerging from the Atlantic suggests some continuity in exchange, but an overall reduction in importation from the Mediterranean in the middle decades of the 5th century (Fernández 2014, 128, 415-30). It is likely that this pattern will have been reflected, and potentially exaggerated, at the northern reaches of this system.