It was quickly observed that the 1990 excavation had exposed the lower edge of the wall in some places and this continued until the bottom edge was exposed all round. In places the wall had been constructed on peat or ash layers which overlay a thick layer of heather. This in turn overlay a substantial deposit of marine molluscs consisting mostly of oyster shells.
The part of the site closest to the shore was higher than the south side which had collapsed substantially in the past. The north side had not collapsed because it was founded on the edge of a substantial stone platform which may have been built to support the dun. The stones of this platform projected up through the layer of heather which was substantially removed in 1990 to expose the upper stones. Habitation material had built up on the platform and slipped over the edge to create a wider, sloping edge to the island.
It is not clear how deeply the platform has sunk into the lochbed silt or how much silt has built up around it since its original construction. However, it is clear that the stones of the platform had been covered by occupation debris and a thick layer of heather and it may be postulated that that debris had spilled out from the edge of the platform so as to convince the later dun builders that there was a sufficiently sound foundation on which to build considerable stone structures.
The sub-circular stone structure (Structure 2) had been built over the edge of the stone platform and because of this the south side of it began to subside very soon after construction. The floor was built up with stones and peat but eventually the walls were pulled down and used for the foundation of a second phase of construction. This in turn also collapsed and eventually the building was dismantled and the stones were used to construct another building closer to the shore and therefore less prone to subsidence. This building (Structure 3) was in turn robbed of its stone to build later structures.
The first floor in the building was covered with a deep layer of peat blocks or turf. The peat in turn was covered with a layer of small cobblestones which were then covered with very white beach sand. The whole floor was covered with the sand and must have appeared very light and clean for at least a short period of time before it was covered in a layer of straw which was then covered successively with more layers. These floors were made up of heather, straw and peat interspersed with, often discontinuous, ash and grit layers. Occasionally there were layers of peat which seemed to have been deliberately laid down possibly to recover the floor when it became too wet.
At this time the entrance to the building was in the centre of the north wall, directly facing the outer dun wall, and there were no internal walls.
The building was constructed, according to the finds, because the dun-dwellers had need of a workshop. Bone and antler tools and wooden utensils were all discovered in the layers above the sandy floor. Some of the objects are obvious. Scoops need little explanation for function but what were they scooping? It is possible that they were used in dairying for scooping cream. Antler tools may have been used in leatherworking. Weaving is indicated by the discovery of a whalebone weaving comb and spinning by spindles and spindle whorls. Many small wooden toggles and other wooden objects indicate functions which are not known but it is clear that a variety of different functions were carried out in the building.
Eventually, the building subsided at the south side which projected out into the loch. This may have been going on slowly, but constantly, over a period of time until it became untenable. The walls cracked at the sides and the outer wall slipped back (south). By the depth of material on that side it would appear that the subsidence was considerable and that serious rebuilding was necessary.
A number of other features were added to the building during the phase II occupation:
The trough was constructed at a late phase in the habitation of the house. Its function is not clear, although it may have been a store for produce or equipment or it may have been a trough for feeding animals. The surrounding material of straw with significant amounts of animal dung suggests the building was used as a byre at this time so any of the above purposes would be reasonable.
At about the same time as the construction of the trough the floor between it and the doorway was paved with natural stone slabs. The floor to the right of the door was also paved with slabs. These had been covered by substantial layer of ash, presumably from the small hearth at a higher level when the building had yet again been reconstructed and possibly related to the overlying wall remains.
The building was less a workshop at this time, was certainly a stock byre and probably also a peat storage area, according to the amount of peat (including cut peat blocks) in the floors and the embedded peat basket. Various small finds indicate that the building was still occasionally used for craft industry.
They were mixed farmers with crops and animals. There may be some doubt as to whether they themselves grew crops or bought in grain and straw since the immediate surroundings of Loch Bharabhat are not suitable for cereal cultivation. Straw is found throughout the occupation layers and the remains of two rotary querns show that grain was being ground on the site. They certainly kept animals as animal dung is common in Structure 2 and spores point to the same conclusion. Numerous dung beetles were also recovered from the floor deposits. The bones of cattle, sheep and pigs are representative of domesticated animals.
The surrounding hills and moors supplied them with tools and food from the red deer that they hunted. It is clear that they hunted the deer deliberately as the antler found on the site has been torn from the skull and not collected from the hillside after shedding. The proportion of deer bones on the site, many with butchery marks, is comparable to those of cattle, sheep and pigs so deer were an important source of meat. It would seem that they also kept dogs as the ends of long bones in some cases have the teeth marks and damage of having been gnawed and it is interesting to speculate whether they were hunting dogs.
The nearby marine environment was important to them. They gathered shells from the shore, such as limpets and winkles, but also exploited the deeper water for oysters and scallops. Fish were also caught as bones from various sizes were found among the ash from the hearths. Other bones show that they also killed seals and a whalebone weaving comb shows possible exploitation of stranded whales. They even used sand from the beach to make a clean floor in the workshop.
Their immediate surroundings also supplied them with fuel for the fire in the form of peat. That they cut and gathered peat is clearly indicated by the amount of peat on the site in the form of cut and burnt blocks and most clearly by the peat basket discovered embedded in the floor. The basket is very similar to those used recently in the Hebrides and is a clear indicator of cultural continuity in the area.
Craft industry is indicated by a range of tools and implements made in wood, antler and bone. Two antler awls may have been used in leather working and a two-tined antler implement looks like a small pitchfork. All have evidence of wear on the ends. Another antler object was clearly functional but the function remains unknown. A wide range of wooden objects were recovered. Two wooden scoops, one rough and one smooth, are in the shape of scallop shells and may have been used for scooping cream in dairying. Another wooden scoop with a well-formed handle looks more like a measure of some sort, possibly for grain. A number of toggles with a range of features, half-checks and holes, show obvious similarities but their function is not clear and the same applies to other complex wooden objects.