234 THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 
since many analyses (pollen and mollusca) show 
that the outer fringes of chalkands remained to 
some extent wooded until the Bronze Age, and 
there is widespread evidence for a phase of later 
Neolithic woodland regeneration (Scaife 1988; 
Evans 1992). It is concluded that Durrington: 2 
could easily be correlated with this extensive 
evidence of Neolithic activity in the area. If so, the 
pollen evidence clearly indicates the prevalence of 
grassland which may be attributed to pasture, and 
also cereal cropping both of which held scrub 
colonisation at bay. It is likely that extensive 
woodland clearance resulted in locally high water 
tables, reduced evapotranspiration and increased 
surface run-off, all of which contributed to the re- 
initiation of sedimentation in the Avon valley. 
The temporal span of zone Durrington: 2 is 
unknown. It possibly spans only middle and late 
Neolithic activity, although it seems more plausible 
that continued land use into the Bronze Age was 
maintaining conditions in which peat accumulated 
on the floodplain. The major change in floodplain 
vegetation from open, wet sedge fen communities to 
drier alder carr is interesting since this apparently 
occurred during the Roman (or post Roman) 
period, representing a period of drier floodplain 
conditions which allowed the succession of carr 
woodland. This would indicate that the floodplain 
had standing water for only two or three months of 
the winter. This reverted to sedge fen but with 
meadow and fen herbs dominated by Filipendula 
ulmaria (Meadowsweet). This change may have 
been through natural causes or by clearance of the 
valley carr wood. 
FOLLY BOTTOM 
by Michael F. Allen 
The pipeline trench traversed the large dry valley of 
Folly Bottom, incised into the Middle Chalk of 
Salisbury Plain to the north-west of Amesbury. 
Only a shallow colluvial profile was revealed and 
comprised a gravel fan in a weakly calcareous dark 
silty clay matrix which overlay late Pleistocene/ 
Devensian Chalk meltwater deposits and sealed a 
relict tree hollow containing a reddish-brown silty 
clay loam. 
Stratigraphy 
An irregular pocket of dark, reddish-brown, 
mottled silty clay loam with occasional flint 
nodules was recorded beneath the colluvium, 134 
(Site 8, Figure &). It contained charcoal flecks 
throughout and was possibly a tree-throw hollow 
which contained evidence of a relict mature palaeo- 
argillic soil. It was sealed by a gravel fan comprised 
of medium to large flint nodules in a dark silty loam 
matrix (133) situated on the edge of the valley floor 
(Allen 1992, fig. 4.3 and cf Allen 1988, fig. 6.5) and 
which originated from valley side erosion. A thin 
silty, stonefree calcareous layer (174) sealed the 
gravel fan, but terminated downslope. This 
probably represents the erosion of fine-grained 
material, possibly as a slurry. 
The section was carefully cleaned but no 
artefacts were recovered. A series of samples was, 
however, taken for molluscan analysis, but 
produced very few shells; all species were typically 
open country. 
Discussion 
The basal tree-throw hollow indicates the presence 
of former argillic brown earths/brown earths and 
the presence of charcoal may indicate deliberate 
felling perhaps associated with the Neolithic 
barrow of Longbarrow Clump on Bulford Down. 
The overlying flint gravel horizon indicates high 
energy erosion. Augering showed that this deposit 
extended for almost 80m along the axis of the valley, 
as well as down the valley side. This may, therefore, 
represent the coarse channel deposit of a temporary 
winterbourne, or high energy erosion down the 
valley axis (Bell and Boardman 1992) This erosion 
was probably responsible for truncating and 
stripping out any previous, possibly prehistoric, 
colluvial deposits which may have been transported 
further down the valley axis. By analogy with other 
colluvial deposits in Wessex (Allen 1992) it is 
plausible that this belongs to the later prehistoric 
period. The lack of colluvium does not therefore 
necessarily represent a general lack of erosion and 
long term land-use. 
PART 3: EARLS FARM 
DOWN (SITES 9-19) 
ARCHAEOLOGICAL 
BACKGROUND 
The palimpsest of archaeological features on Earl’s 
Farm Down (Figure 8) forms part of a wider pattern 
of linear ditches and trackways of Bronze Age, Iron 
Age, and Romano-British date which extends over 
