NEOLITHIC AND LATER PREHISTORIC LANDSCAPE OF THE AVON VALLEY 239 
Table 3. Molluscs from Earl’s Farm Down ditch 3 
Feature Ditch 3 
Gontextey win cin enn ae 20------------ 19/20 19 18 --------------- 5/6 --------------- 4 
Sample 16 1S 14 13 12 11 Os 8) 8 7 6 Si 4 3 2 1 
Depth (cm) 150-160 140-150 130-140 120-130 110-120 100-110 87-100 78-87 70-78 60-70 ~—-50-60 ~—s 40-50 30-40 ~—S 20-30 -~—«*'10-20 0-10 
Wt (g) 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 ©1000 815 
LAND 
Pomatias elegans (Miller) ] 2 - - 1 ] - 4 2 2 2 3 3 4 
Cochlicopa lubrica (Miller) - - - - - 1 1 2 1 - - 3 
Cochlicopa spp. 1 - 1 il 3 4 4 5 3 3 2 3 1 - 
Vertigo pygmaea (Draparnaud) - - - 1 2 3 1 1 2 1 2 
Vertigo moulinsiana (Dupuy). - - - - - - - - 1 - - - - - - - 
Pupilla muscorum (Linnaeus) 68 99 125 73 44-98 192 570 794 741 558 259 141 114 79 60 
Vallonia costata (Miller) 34 337) 162 24 13 43 98 64 79 70 76 28 28 22 14 1 
Vallonia excentrica Sterki 7 14 11 2 14 52 20 46 36 24 9 9 25 20 8 
Vallonia spp. - - - - - 8 7 8 6 7 6 4 - - 
Punctum pygmaeum (Draparnaud)  - - - - - 1 6 | 4 1 1 — 
Vitrina pellucida (Miller) - - - 
Nesovitrea hammonis (Strom) 2 2 
Limacidae 2 2 2 - - 
Ceciltoides acicula (Miller) - 1 - 2 1 2 28 
Cochlodina laminata (Montagu) - 1 1 - - 
Clausilia bidentata (Strom) 2 - 1 - 
Candidula intersecta (Poiret) - 6 1 - 
Cernuella virgata (Da Costa) - - - - - - - - - - - - - 1 
Helicella wala (Linnaeus) 10 5 12 1 3 ll 17) » 40 24 22 20 16 10 19 9 4 
Tnrchia hispida (Linnaeus) 11 16 23 2 4 14 52 14 24 24 14 17 1 2 - 
Ananta arbustorum (Linnaeus) - - - - - - - - - 
Helicigona lapicida (Linnaeus) - - 
Cepaea/Ananta spp. - - - - - - - 2 - - 
Taxa 7 6 6 6 6 6 9 11 10 12 11 8 9 8 9 
TOTAL 132 169 234 103 69 181 432 729 989 919 709 341 205 190 131 81 
assemblage. These are species also found in short- 
turfed grassland. The absence of the shade-loving 
catholic species often found in longer grassland, 
such as Carychium tridentatum, indicates a well- 
established short-turfed (trampled or grazed) 
grassland, and the homogeneity of the local open 
landscape. It may also suggests the absence of long 
vegetation colonising the ditch itself (see Evans 
1972, 322-4). 
The virtual absence of shade-loving species 
indicates that the ditches were cut into a pre- 
existing well established open _ short-turfed 
grassland. The assemblages are both specialised 
and mature ones. The relatively high numbers of 
shells retrieved, together with the absence of 
evidence for long vegetation within the ditches, 
indicate a lack of stabilisation and a constant slow 
process of infilling within the ditches. The 
assemblages therefore seem broadly to represent the 
same general land-use throughout the history of the 
- ditches although there are slight fluctuations 
within the mollusc assemblages (Tables 3 and 4; 
Figure 10). 
Minor fluctuations within the assemblages have 
been attributed to sub-zones within each ditch (see 
Figure 10). These subzones, although based on the 
molluscan assemblages also correspond to the 
tripartite ditch fills (cf. Evans 1972, 322-8; 
Limbrey 1975, 290-300; see Figure 10). 
The subzones seem to reflect localised changes 
in the intensity of land use, particular to grazing. 
The ditches were dug into, and existed in, a short- 
turfed grass downland (sub-zone 1) indicating long 
established open (grazed) downland prior to their 
construction. During the natural sedimentation of 
the ditches, increased, or more intensive grazing 
(possibly even over grazing) and the creation of 
bare patches of soil (Pupilla and Pomatias) is evident 
(sub-zone 2). Finally, grazing pressure is reduced 
and a slightly longer grassland sward established, 
but with hints of localised and _ possibly 
intermittent arable activity (sub-zone 3). This 
occurs from at least the medieval period and later 
and may be compatible with the establishment and 
use of the Romano-British field systems visible on 
aerial photographs. 
Conclusion 
The land-use of the surrounding areas seems to be 
one of open pasture throughout the history of the 
ditches with a little arable activity coming in late 
on. The paucity of shade-loving species from these 
sequences indicates that not only had clearance 
occurred some considerable time prior to ditch 
infilling, but also that the grazed downland was 
long established. This would therefore not be 
