EXCAVATIONS IN 1999 ON LAND ADJACENT TO WAYSIDE FARM, DEVIZES 



161 



British date, although a number of pieces of 

 prehistoric worked flint was also present. 



The midden 



Across a large part of the central portion of Area 2, 

 and in parts sealing the buried soils described above, 

 was an extensive deposit interpreted as a midden 

 (overall contexts 4255 and 4293). The extent of 

 the deposit can be seen on Figure 7 and it is also 

 shown on Plates 4 and 5. The midden was on a 

 broadly north to south alignment, covering an area 

 of c. 1150m 2 . The material comprised an 

 homogenous dark brown/black humic sandy clay, 

 undulating across the area with a depth varying 

 between 0.3m towards the centre and 0.1m on its 

 fringes (overall average was 0.15m). It is likely that 

 the midden was formed as a result of a single episode 

 of dumping, as identical soils (and indeed conjoining 

 pottery sherds) were present in the adjacent ditches 

 and pit F4225. The deposit contained the vast 

 majority of artefacts recovered from the site, most 

 of which are of late Romano-British date. The 

 southern extent of this deposit was located within 

 the excavation area, although it was not fully revealed 

 to the north. The midden soils appear to widen 

 towards its southern end, following the area denned 

 the two ditches (F4261 and F4254) which also 

 become further apart at this point. 



PHASE 3: POST-MEDIEVAL 



A single ditch can be allocated to this phase. F4004 

 ('modern field boundary' on Figure 2; section on 

 Figure 8a) was aligned approximately north-south 

 with a steep-sided profile. Its alignment suggests a 

 former continuation of an existing boundary. Nine 

 pieces of blue transfer-printed wares, white 

 frnewares and stonewares of 19th-and 20th-century 

 date were recovered from the fill (context 4005). 

 Four clay pipe stems and a fragment of modern 

 clear window glass were also present within the fill. 



TRENCH RESULTS 



Five trenches were excavated, radiating to the east 

 and south of the main excavation Area 2 (shown 

 on Figure 2 as Trenches 19, 21-24). Trenches 19, 

 21 and 22 contained evidence for archaeological 

 activity. Despite scanning of all spoil heaps adjacent 

 to the negative trenches there was a complete 

 absence of artefacts of any date. 



Trench 19 revealed evidence for the 

 continuation of ditch 4075. This was visible only 

 below subsoil layer (4074) although this relationship 

 was unclear due to the similarity of the fill with the 

 overlying layer. 



A surviving buried soil horizon (2105), 0.15m 

 thick, was encountered in the northern half of 

 Trench 2 1 , and appeared to be identical to buried 

 soil (4206) encountered in the main excavation area. 

 This soil horizon (2105) was cut by a single 

 archaeological feature, F2109, comprising a 

 northeast-southwest aligned short, irregular linear, 

 of which c.6m was exposed. 0.5m wide, becoming 

 wider at its terminal, and 0. 1 5m deep, late Romano- 

 British pottery was recovered from its fill The ditch 

 terminated at its northeast end within the trench. 



A single cut feature was present in Trench 22 

 (F2202 on Figure 2) which was cutting a deep 

 subsoil (context 4074) present elsewhere on the site. 

 F2202 was a north-south aligned probable ditch, 

 with a width of 1.50m and steep profile to a depth 

 of 0.5m. This feature probably terminates 

 immediately to the north, as it did not appear in 

 the main excavation area. No artefacts were 

 recovered. 



THE FINDS 



THE IRON AGE AND ROMAN 

 COINS 



by Mark Corney 



48 coins were submitted for identification, dating 

 and comment. With the notable exception of SF233, 

 a late Iron Age silver unit from a late Roman context 

 (4205), all are Roman and, bar one late 3rd-century 

 piece, (SF249) all are of 4th-century date. 



Coin SF233, context 4205, is a fragmentary late 

 Iron Age uninscribed silver unit (VA 1042-1) of a 

 general type Robinson (1977) has described as 'sub- 

 Dobunnic', but now regarded as part of the main 

 Dobunnic series (van Arsdell 1989; 1994). Its 

 presence is of considerable intrinsic interest as it is 

 close to the current southern limit of the known 

 distribution of the type. 



For meaningful statistical analysis, the 

 assemblage has been treated in accordance with the 

 methodology established by Casey (1974) and 

 Reece (1991, 1995). When expressed as per mills 

 totals (Table 1 ) the Wayside Farm group shows 



