Cardiff Studies in Archaeology - Specialist Report Number 18


next page

The excavation strategy for mound 2 was fairly straightforward. The object was to fully define the structure and to excavate the secondary floor layers. The 1999 trench was therefore reopened and the backfill removed (partially by machine). The trench was extended 8 metres to the east, in the belief that this was a generous expansion that would finally reveal the east wall. This extension almost immediately revealed the presence of a house (Figure 2 and 3), oriented north to south, whose west wall lay parallel to and only some 0.30 m from the east end of the 1999 trench!

The presence of this house restricted the amount of work that could be done and required further small extensions to the original trench. It was originally expected that we would be able to remove the deposits infilling the original hall in the first four weeks. In previous years these had proved to be relatively sterile sand layers. The presence of a house meant that this was not possible; floor levels and secondary remodelling (very obviously present) would require careful excavation. It was also felt that precipitate removal of structural elements, prior to the full exposure of the house, was not justified. However, it was still very important to define the east end of the underlying hall and it was decided to excavate a 2 m trench across the centre of the later building outside the area of obvious secondary modification. The trench was positioned to continue the longitudinal section through the early hall. The trench was also extended to the west to enable the dismantling of the unstable west wall of the hall and small test trenches were dug, north and south of the east end, to define the full extent of the later building. By the end of the excavation a total area of 25 m by 7.5 m was open.

The most surprising and important discovery made this year concerns the floor layers inside the hall. Last year these deposits were split into primary and secondary floor layers (Sharples 1999, 14 - see also the on-line 1999 interim report). This year the dismantling of the collapsed west wall revealed that the "primary floor" extends over 3.80 m to the west and belongs to an earlier timber structure. The floor was only partially explored but was defined to the north by a line of post holes in a gully and has a line of post holes running down the interior. Excavation of the floor deposits and the overlying sands resulted in the discovery of several fragments of steatite vessel and not much pottery.

After the abandonment of this house the primary floor was covered by layers of relatively clean sand. Some of these layers may be naturally accumulated wind blown sand but others were clearly dumped, though the source was relatively sterile sand. These deposits underlay the west wall and appeared to represent deliberate infilling and covering of the original house prior to the construction of the stone structure that was associated with the "secondary floor".

This structure has now been defined as a bow sided hall 19.3 m long and about 5.8 m wide (Figure 4 and 5). This year saw the excavation of the western eight metres of the floor of this hall and has exposed a further 5 m. The inside wall face of the east end of the house was defined in a 2 m slot through the floor of the later house. The principal feature of the floor was a hearth running down the centre of the house. This was up to 1.50 m wide and is over 9.10 m long. The hearth was surrounded by various occupation deposits and was covered by a final layer of dark grey sand.

The most striking feature of these floor deposits has been the presence of a large number of pits (Figure 5). At the west end of the structure a total of 15 pits were found in the last four metres of the floor and some may have been missed in the original sondage. There is also a very substantial pit immediately below the central hearth. These pits vary in size shape and fill. The earliest are shallow scoops and are found cutting the "primary floor". The later pits tend to be deeper steep sided features, which were dug immediately prior to the accumulation of the final floor, after the house had been occupied for some time. Very few of these features had anything distinctive in their lower fills to indicate why they had been dug. If finds were present they were in the upper layers and seem to be the result of subsidence from the overlying floor layer. This layer continued to produce an extraordinarily rich and varied material culture which this year included an extremely important cylinder decorated in the Ringerike style (see below).

Sitting on the "secondary floor" in the western half of the building was a complex arrangement of walls that represent the remodification of the building. These structural remodifications were subsequently truncated by a substantial quarry hollow, which removed there northern half. The complex sequence of modifications and the subsequent modifications make it very difficult to understand this phase of activity but it is similar to the modifications that occurred at Cille Phaedair in phases 5 and 6 (Brennand, Parker Pearson and Smith 1998). There then followed a period of sand accumulation, which precedes the construction of the large north south oriented house.

The later house is a rectangular, straight sided building, 12 m by 5.40 m (Figure 6). The walls survived to a fairly even height of 0.7 m throughout the area excavated this year. There was no evidence for collapse and it seems quite likely that the superstructure to this house was of timber and turf, as has been hypothesised for the house on mound 3. Limited examination of the floor suggested the house might have originally been slightly narrower, 4.10 m, and that the west wall was rebuilt. A central hearth was exposed and showed signs of being replaced at least twice. In the later phase the hearth was surrounded by winkles that seem to be carefully placed to form a kerb. Underneath the hearth there were three shallow pits whose filling all included thin charcoal rich layers. These pits were clearly dug, in sequence, prior to the occupation of the house.

After the abandonment of this house a small structure was created in the south end. This structure utilised the original west and south walls but had new east and north walls. There was an entrance in the north wall that led into the northern half of the original building. The interior was roughly 4 m by 2.90 m. Unlike the secondary structures discussed in previous years (Sharples 1999, 28 - see also the on-line 1999 interim report) this had a distinctive floor which was associated with hearth deposits.

next page