2010 Knowe of Swandro Excavation
Between the 23rd of June and the 14th of July, a team from the University of Bradford, Orkney College (UHI) and City University New York cleaned, recorded and sampled three sites in Rousay: the broch mound at Brough, the ditch at Midhowe and the Knowe of Swandro, a mound on the Bay of Swandro. The site is subject to coastal erosion and close to the Norse house site at Westness.
The objectives for the 2010 season’s work were to locate a reported earlier excavation trench and to characterise the mound, which has been variously described as a broch, a ‘mutilated turf-covered mound’ by the Ordnance Survey in 1967, and as a ‘stony mound’ by the Investigators of the Royal Commission (RCAMS 1946). The planned investigation extended onto the beach rather than along the coastline as it became obvious that the site extended under the storm beach towards the tide line.
Looking roughly south east along Eynhallow Sound over the Bay of Swandro, the Knowe of Swandro marked by a red star. Thanks to Darren Oliver for the photo.
A small trench, 2.5 x 8m, was opened across the SE end of the curving bank and hollow on the top of the mound. The hollow centre of the mound, which had appeared to be the result of earlier investigations, seems not to have been disturbed; there was a fine and even layer of shillet sealing this area which appeared to be a weathering deposit. A Cleaning the erosion face of the South Howe mound tumble of rocks sealed the shillet and butted a stone feature which appeared to be structural and may be a partially destroyed length of wall.
Two other small stone features may also be fragments of walling but this cannot be determined without more extensive excavation. There was an area of paving composed of large, flat worn stones in the E of the trench. This was sealed by a small patch of limpet midden in the NE corner. The very top of an orthostat was visible in the SE corner of the trench. The area was cleaned, planned and photographed and the contexts recorded. A number of conclusions can be drawn on the basis of the evidence from this season’s work. There are a number of phases to this part of the mound and the presence of worn paving at the very top of the sequence, sealing or butting an earlier wall, suggests that the site is composed of more than a single structure. The presence of the undisturbed shillet suggests that wherever the earlier investigations were located, the hollow area is not the result of this but rather represents a weathering layer over undisturbed contexts.
Orthostats (upright stones) poking through the beach deposits indicating buried structures
The tops of a series of orthostats had been noted among the boulders and shingle of the storm beach and appeared to be a previously unrecognised part of the site. Investigation of these features completely changed the interpretation of the mound. An area, c.5 x 6m, stretching from the erosion bank at the top of the beach down towards the sea was cleared. The orthostats, which had appeared level with the boulders and shingle, survived to a height of at least 0.5m and appeared to form the backs and sides of three cells of a curving dry stone structure, whose projected centre was somewhere below the current high tide mark. The largest cell, to the E of the feature, contained several phases of paving and also had paving in front of a long dressed kerb stone which formed the front of the cell. The two smaller cells to the W also retained areas of paving and patches of ash rich midden survived between and on top of the stones. This midden produced well preserved bone and pottery which on initial examination appears to be of Late Bronze Age or Early Iron Age date. The structure appears to continue towards the high tide mark and lies well within the area of the spring tides and of storm events. The back of the structure was sealed by more midden of a later date, which appeared to be earlier than the features investigated at the top of the mound, although this cannot be confirmed without further investigation. Bulk samples were taken from the midden contexts within the structure. The charred plant remains and animal bones will provide palaeoeconomic information and radiocarbon dates.
The excavated beach area was consolidated with geotextile, sandbags and boulders, whilst the area at the top of the mound was backfilled and returfed.