By Patrick Foster
Shiant Islands Project (SHIP) 2002
Click on any photo to see a larger version.
A third season of work was carried out on the islands during four weeks in June. Excavation continued on House Island at the Blackhouse Site HI 15 (NG 4199 9726) where pre-blackhouse deposits and structures were revealed in greater detail. The blackhouse itself is of traditional form with combined human and animal occupation under a single roof. There is a winnowing barn on the southern side of the house, which may be a later addition. Several definite structural modifications, which include an additional barn added to the northern side of the house, suggest a relatively long period of use. The factory-manufactured pottery is currently being analysed, but preliminary dating indicates a range between 1770 and 1900. In the lower deposits of this period hand made local pottery predominates.
Beneath the blackhouse and its ancillary buildings a more complex structure composed of several rooms and hearths has been partially revealed. Several walls continue out of the excavation area suggesting that the underlying building is not only more complex in plan but is also considerably larger in extent. Pottery from the pre-blackhouse deposits is entirely composed of handmade 'craggan' type ware with a number of finer jars with 'S' shaped hook rims that may indicate a late medieval or early post medieval date. A second steatite spindle whorl was found.
Also on House Island an injured member of the team managed to excavate the interior of one of the early 20th century lobster fishermen's huts (Site HI 4)(NG 4186 9767), which possessed beach cobble flooring and raised bench seating around an open fireplace against the northern wall.
At Annat on Rough Island one of the three circular stone huts (Site RI 41b)(NG 4117 9829), which was barely visible at ground surface level and had only been discovered by an evaluation sondage last year, was opened. Below the turf the interior of the hut was filled with a mass of displaced walling and roofing stone in which no collapse pattern could be seen. This is not surprising since a large amount of material has obviously been robbed from the site to construct the modern sheep fank and its associated enclosure walls.
Beneath this upper destruction layer a second deposit of collapsed stonework was identified as a change in the peat soil, to a darker, denser and more compact matrix . This deposit had apparently accumulated above a compact deposit of post-abandonment peat soil the surface of which had been used as a floor level, although it was very uneven with many protruding stones. Confirmation that the abandoned hut had been re-occupied, albeit on a temporary or sporadic basis, was found in an unstructured hearth composed of a large mound of peat ash against the east wall of the hut. The east entrance passage appears to have been deliberately blocked at this time. Pottery found associated with this hearth and 'floor' level was composed largely of 'craggan' ware. This late period of re-occupation of the decayed building is interpreted at present as being a shieling of the 18th to early 19th century.
Below these late contexts the first internal structural details began to emerge. Although even at this lower level rat burrowing had still managed to displace some of the stonework a pattern of stone slabs set on edge could be seen to form internal divisions. Firstly a large square central fireplace set square on to the east entrance, which was now fully revealed as a downward sloping passage into the interior of the hut. Radiating from each corner of the fireplace further slabs divided the internal area of the hut into four compartments. The southern compartment was made more complex with a channel around the inside of the outer wall. Possible further structural complexities made with clay were just visible.
The central fireplace was filled with a slightly mounded deposit of peat ash, some of which has been trampled out into the entrance passage. A black greasy peat soil appears to fill the rest of the compartments and there are two further possible small hearths in the western and southern compartments.
Several hundred sherds of pottery were recovered from the surface of the deposits. Most were in discrete groups and have been found to refit forming several almost complete profiles. The form appears to be a standard flat-bottomed bucket-shaped jar standing up to 0.20m high. The rim diameters are, more certainly, between 0.20 and 0.24m. The fabric is of local clay in which the mineral inclusions have apparently received the minimum of refinement and being fired at a low temperature the vessel produced is of an extremely friable nature. Thick carbonised deposits of peat soil, caking the external surfaces, are almost certainly acquired in the initial firing in the domestic hearth and may indicate that most of these large fragments are from failed firings. Also many sherds display extremely rough-cast external surfaces which do not appear to show any wear. One large rim and body fragment however has a repair hole drilled after firing which must indicate a period of normal use.
At present secure dating must await a suitable C14 sample, but a suggestion that these wares are Pictish (pers. com. Mike Parker Pearson) conforms with the author's own impression.
At present the site appears to be composed of a massive rubble stone platform, possibly a natural landslip, approximately 2m high at the leading edge, against the hillside of a small inlet on the south coast of Garbh Eilean or Rough Island. On the platform it has been possible to identify three stone-built circular huts, but the presence of modern structures may mask the presence of others. The huts each have a diameter of just over 3.5m internally. The internal face of the huts is constructed with well placed blocks of local stone, but the outer face appears to be insubstantial and was most likely of mixed turf and stone, the core being rubble and earth. In the excavations many large flat elongated slabs indicate that the roofing was of stone corbel construction and the huts would have been beehive shaped. The floor areas appear to be slightly sunken with the east facing entrance passages gently sloping down into the interior.
A 2m wide stone wall encloses a small part of the steep slope to the north of the platform and also includes the platform. A small sondage against the outer eastern face of the enclosure wall showed it to be very well constructed of large naturally square shaped stone blocks. A sondage cut last year showed that such a face was also present on the interior and that the core was of earth and rubble. This type of enclosure construction is not reproduced at any other place in the islands.
The site has an annat place name attached to it, the exact meaning of which is complex, applied either to old church sites, or to abandoned church sites, or to sites which at some time belonged to the church.
The discovery of an early Celtic Christian carved cross stone in the blackhouse excavations in 2000 has strengthened the feeling that an early Christian site, often alluded to in Shiant folklore, exists somewhere on the three islands. The typology of site RI 41b, its beehive huts and elaborate enclosure, the annat place name and the suggested date range of the pottery are all arguments supporting the possibility for an early Christian site on Garbh Eilean. Even if this proves to be in error the discovery of a domestic site of Pictish date is still of great importance to the archaeology and history of the islands.
Finally, a short excursion over to Mary Island allowed an examination of the possible chapel site MI 1 (NG 4308 9860) and it is possible to confirm that the rectangular stone structure is constructed with a locally made lime mortar that contains sea shell and small stone inclusions. A sample was taken for analysis.