188 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 



Table 9: Comparative figures for midden and non-midden groups expressed as a percentage of total EVE's 



OVWWH HARSH 



OXF 



NFCC 



OXFWH 



DOR BBI 



GS 1 &2 



OXID 



AH 



RS,PH 











SI &S2 





Midden 40.7 



10.5 



2.6 



14.8 



19.7 



5.6 



1.6 



Non-midden 27.5 



2.5 



1.8 



20.8 



41.3 



5.6 



0.4 



2.6 



1.2 



Decoration 



Decoration on the coarse wares is largely restricted 

 to burnished obtuse lattice on the central body zone 

 of type 3 jars; rilling on the central body zone of 

 hooked-rim jars in fabrics OVHWH, HAR SH and 

 OXID S2; white or dark grey slip over the rim of 

 jars in fabric AH RE and random curvilinear 

 burnished patterning on the base of bowls/dishes 

 in forms 20 and 25 in DOR BBI. 



MARKETING AND POTTERY 

 SUPPLY AT WAYSIDE FARM IN 

 THE LATER FOURTH AND 

 EARLY FIFTH CENTURIES 



The late Roman pottery assemblage from Wayside 

 Farm can be confidently dated to the early 5th 

 century (see below), and as such provides a rare 

 insight into regional ceramic supply patterns at the 

 very end of the Romano-British period. The 

 character of the site is somewhat problematic given 

 that the excavation examined only part (and 

 probably only the periphery) of a much more 

 extensive complex. 



The most immediate pattern is the very large 

 proportion of Oxford fine wares present. On the 

 midden site these represent over 40% of the EVE 

 total (Figure 23) and almost 30% by weight (Table 

 9). Added to this are the small number of shell- 

 tempered products from the South Midlands 

 ( 1 .23% of the EVE total), Alice Holt/Overwey kilns 

 (2.61% of the EVE total, see Figure 23) and Oxford 

 white-ware mortaria (2.6% by EVE). Taken 

 together, production centres over 50km to the 

 north-east and east ofWayside Farm were supplymg 

 almost 49% by EVE of the ceramics to the site. 

 Local supplies, exclusively coarse wares with grey 

 sandy products dominating, account for 25.5% by 

 EVE of the supplies. The remaining ceramic 

 categories (25.3% by EVE) can be sourced to 

 production centres over 50km to the south of 

 Wayside Farm in the New Forest and the Poole 

 Harbour region (Figure 23). It is immediately 



apparent that nearly 75% of the pottery deposited 

 in the midden at Wayside Farm had travelled a 

 considerable distance to the site. 



The patterns for non-midden deposits are 

 slightly different, with a higher percentage by EVE 

 of local coarse wares (46.9%) although the strong 

 links to the north-east (Oxford region) and south 

 (New Forest and Poole Harbour area) are still 

 marked. The higher percentage in local supplies in 

 non-midden deposits may be chronological, or, 

 functional (see below). The variation between 

 midden and non-midden supplies is presented 

 above in Table 9. 



Detailed understanding of these patterns is 

 hampered by the lack of comparable, well-excavated 

 and quantified assemblages from nearby rural sites 

 and the local 'small towns' that must have acted as 

 marketing centres. The nearest known 'small towns' 

 are at Cuneao-Mildenhall (Corney 1997, 2001), 

 some 20km to the east, and Ver/ucio-Sandy Lane, 

 some 9km to the north-west. No Roman road link 

 to either site is known in the Devizes area although 

 such may reasonably be expected. 



The pattern as discussed shows a very high 

 proportion of pottery supply over 50km from the 

 site. This pattern is beginning to recognised over a 

 wide area of lowland Roman Britain, and is seen as 

 indicative of a few production centres continuing 

 to supply pottery after the collapse of local industries 

 in the late 4th and early 5th century. The matter is 

 discussed further below. 



DISCUSSION, DATE AND 

 CONCLUSIONS 



The late Roman pottery assemblage from Wayside 

 Farm is an important group for the region. The high 

 proportion of fine wares from the Oxford region that 

 can be relatively closely dated demonstrate that 

 accumulation and deposition must have occurred 

 after C.AD350/360. Indeed the number of distinctive 

 late decorated hemispherical bowls with rosette and 

 demi-rosette decoration may indicate a post AD370 

 date and the presence of such forms in association 



