POOR HILL 45 
the terrace was cut into the slope to form a lowish scarp and the front (eastern) surface 
appeared to have been built up with the resultant spoil to extend the platform over the 
slope. The back drain led into side ones which became “V” shaped depressions 
directing the water to the amphitheatre gully. 
Except for a very thin layer and a scrappy litter layer the surface of the ground 
proved to be the original floor of the terrace. This produced little except large smudges 
scattered about the excavation. It was decided to follow one of these down into the soil 
profile. At an approximate depth of 30-40 cm the smudges. turned into beautifully 
defined postholes where they entered the hard clay pan. By vigorous trenching the 
posthole line was followed, by-passing the confusing smudges on the way. It produced 
the layout of posts as shown on the terrace plan (Fig. 4). Gorbey suggests that, due to 
the acid(?) soil environment, the postholes were not clearly evident until from 30-50 
cm down. Their upper profiles had been removed by chemical and soil action. 
The posts appeared generally large and rectangular. A few postholes were fully 
excavated and, of these, a number were clearly slabs and buried quite deep (no details 
of measurement or depth are available). The posthole distribution suggested a large 
structure some 15 x 7 m in size. 
A cluster of rocks and charcoal was found inside the structure against the south- 
west wall near posthole “P”, Gorbey interpreted this heap as a fireplace but said it was 
not a four sided hearth. No entrance way was established although the presence of a 
small posthole “ag” north of but in line with the west wall could possibly suggest the 
presence of a light porch. 
Parts of the interior of the structure were tested but only the possible postholes 
“ai” and “ah” gave any indication of some sort of roof support. There was no evidence 
for a large centre back post to support a ridge pole. Testing confirmed the surface 
evidence of the drain running along the back and sides of the terrace. Only a few flakes 
of obsidian and chert were found. These are not now available. 
A smaller, now unidentified, terrace was tested by Trevor Hanson at the same 
time. Two sides of what appeared to be a small rectangular structure were located 
(Gorby pers comm.). There seemed to have been a replacement of posts so that the 
structure could have stood for some time. A core (stone unspecified) with several 
fitting flakes came from this terrace. 
At the time of the original survey one of the upper semi-detached terraces (Fig. 
5) on the eastern slopes below the pa (not an amphitheatre terrace) was also tested by 
Davidson, Only the north-eastern third was excavated. A very shallow deposit of top 
soil and litter was removed to reveal an occupation surface and a line of insubstantial 
postholes along the front of the terrace, There were also indications for scattered 
stakeholes and a diffuse semicircular burnt patch containing charcoal showed on the 
surface. This was centrally placed on the terrace but only half was exposed by the 
excavation. It was about 80 cm in size. 
