EXCAVATIONS EST 1999 ON LAND ADJACENT TO WAYSIDE FARM, DEVIZES 



189 



with scattered coin hoard of post AD370 date from 

 Cunetio adds weight to this (Moorhead 1997). The 

 presence of products from the Alice Holt/Overwey 

 kiln complex and the South Midlands, probably 

 Harrold in Bedfordshire, may indicate an even later 

 date, quite possibly into the first quarter of the 5th 

 century. Certainly there is growing evidence for the 

 longevity of the late Roman repertoire from the Alice 

 Holt and South Midlands production centres with 

 products reaching increasingly distant centres (cf 

 Going 1988, 70-71, in discussing changing trends 

 in coarse ware production and supply in early 5th- 

 century Great Dunmow, Essex). Further and 

 ongoing work in Essex and other parts of the country 

 (Lyne pers. comm.) is beginning to show certain 

 trends where the occurrence of Oxford colour coat 

 products in association with Alice Holt/Overwey and 

 shell-tempered wares may be a significant indicator 

 of 5th century activity. It is argued that the coarse 

 wares in such assemblages begin to occur further 

 from their source as the established local suppliers 

 begin to decline (SGRP, Newsletter 29) . The highly 

 regional nature of late Roman Britain makes direct 

 comparisons with eastern sites tentative, however the 

 general similarity of the pattern described with 

 Wayside Farm is significant and, in the absence of 

 comparable assemblages nearby, is noted here as of 

 considerable potential importance. The numismatic 

 and other special finds evidence from Wayside Farm 

 supports an early 5th-century date for the midden 

 deposits. Sixteen coins were recovered from the 

 midden and associated pit, F4255. All are 4th- 

 century in date and include six Valentinianic issues 

 of AD 364-78 and a moderately worn issue of 

 Arcadius, minted AD392-402. It is argued that the 

 overall composition of the late Roman assemblage 

 is quantifiably internally consistent and that activity 

 and deposition is restricted to a date range of 

 c.AD370-420+.The ceramic group from the midden 

 pit F4255 (Figures 19-21) is of especial note, 

 containing the latest dateable coarse wares from the 

 South Midlands, Oxfordshire and the Alice Holt/ 

 Farnham region. 



The non-midden deposits, as noted above, show 

 slightly differing clusterings and proportions of 

 fabric types. Overall these differences do not appear 

 to be statistically significant and the main focus of 

 late Roman activity in these deposits is still unlikely 

 to have commenced before the middle of the 4th 

 century. The higher proportions of fine wares in 

 the midden deposit are more likely to indicate 

 specific functional and status differences in the 

 discard patterns on the site. 



Other recently excavated and published late 

 Romano-British assemblages from Wiltshire show 

 strikingly different patterns. At the extensive rural 

 settlement on Butterfield Down, between Amesbury 

 and Boscombe Down Airfield (Rawlings and 

 Fitzpatrick 1996), late fine wares only accounted 

 for 5.2% (by weight) of the total ceramic 

 assemblage, these being almost evenly divided 

 between Oxford and New Forest products. 



At Figheldean, a probable villa and associated 

 settlement (Graham and Newman 1993), the 

 proportions are similar to those on Butterfield 

 Down. Late fine wares represent only 5.2% (by 

 weight) of the ceramic assemblage with Oxford 

 products slightly better represented than those from 

 the New Forest. 



Whilst the late date of the Wayside Farm 

 assemblage appears secure, the character of the site 

 is more problematic. The absence of significant 

 quantities of mortaria could, on its own, be seen as 

 a chronological trait and support a very late date. 

 However, the further absence of large storage jars 

 and other 'everyday' domestic type vessels such as 

 colanders, coupled with the high percentage of fine 

 wares may also indicate a more specialised activity 

 on or near the site. The unusual nature of many of 

 the deposits recovered from pit F4255 (cf Bircher; 

 Mills this volume) may be of a ritual character and 

 it is conceivable that the excavated area is adjacent 

 to a more specialised focus, perhaps a shrine. The 

 true nature of the whole site can only be ascertained 

 through further fieldwork. The ceramic assemblage 

 however is of undeniable regional importance and 

 the dating evidence of ceramics, coins and other 

 special finds clearly marks the deposit as one that 

 belongs to the very end of the Romano-British 

 period. 



Catalogue of illustrated late Romano- 

 British pottery 



Fig. 19.1. Context 4226 (fill of pit F4225). Fabric HAR 

 SH. Hooked rim jar with horizontal rilling over body of 

 vessel. 



Fig. 19.2. Context 4226 (fill of pit F4225). Fabric 

 OVW WH. Hooked rim jar with horizontal rilling over 

 body of vessel and pale cream slip over external surface. 

 Lyne and Jeffries form 3C. 



Fig. 19.3. Context 4226 (fill of pit F4225). Fabric 

 OVW WH. Hooked rim jar with horizontal rilling over 

 body of vessel and pale cream slip over external surface. 

 Lyne and Jeffries form 3C. 



Fig. 19.4. Context 4239 (fill of pit F4225). Fabric 

 GS1. Lid. 



Fig. 19.5. Context 4226 (fill of pit F4225). Fabric 



