St Kilda, the little group of islands far offshore from northwest Scotland, was known for its seabird subsistence in the period before its evacuation in 1930. Recent discoveries suggest that the importance of agriculture in the prehistoric period (before the 16th century AD) may have been underestimated.
St Kilda and its prehistory
The St Kilda archipelago lies some 60 km west of the Western Isles of Scotland (57 [degrees] 49 [minutes] N; 08 [degrees] 35 [minutes] W). Apparently only the main island, Hirta, has been continuously occupied; its population of 36 was evacuated in 1930, much reduced from its highest recorded figure of 180 at the time of Martin Martin's visit in 1697 (Stell & Harman 1988: 31). The writings of Martin (1698) and various subsequent written accounts and photographs have revealed quite a lot about the islanders' lifeways over the period 1700-1930 (Macgregor 1960; Steel 1975; Maclean 1977). The archaeology of the area most sheltered from the Atlantic gales, around Village Bay, has been well documented in a recent RCAHMS Scotland volume (Stell & Harman 1988). The landscape is dominated by the ruins of 19th-century buildings, many of them along a 'street', the axis of the village replanned and relocated in 1830. The people rented allotments within the head dyke which was constructed then. There are also numerous cleitean (anglicized as 'cleits'), corbelled stone and turf buildings used mainly for storage. Many were in place by 1697 (Martin 1698: 59). There are earlier walls and banks, notably north of the 19th-century houses, in the areas just above and below the 1830 head dyke (Stell & Harman 1988).
The most up-to-date account of St Kilda's prehistory (effectively the period before the mid 16th century) has to rely on evidence from individual sites and finds (Cottam 1979). There are some Norse place-names, and Martin records that three churches once existed in the Village Bay area; two inscribed crosses are incorporated in standing buildings (Harman 1976-7). There is a souterrain, which should date from probably relatively late in the Iron Age (defined broadly, in Scottish terms, as the millennium centred on the BC/AD transition). Numerous 'boat-shaped' stone settings above the 1830 head dyke have been tentatively ascribed to the Bronze Age (Cottam 1979: 46-8). The island of Dun is said to contain the ancient wall of what may have been a prehistoric 'defensive' site. This article presents three discoveries made in August 1994, when I was a member of a conservation work party organized by the National Trust for Scotland, present owners of these islands; they may change current perceptions of St Kilda's prehistory.
Stone tools in the cleits
The museum in the re-roofed House 3 contains three or four stone tools, including one described (if memory serves) as a possible blade for a hoe or digging implement. Shortly after seeing this, I noticed a lenticular-sectioned piece of stone protruding from the lower part of the turf roof of cleit 55. I picked it out, and saw that it had once been part of a stone implement. A search of cleits in the vicinity soon produced two more broken implements, this time from among the stonework. Speculating that a more systematic search of the cleits might produce a serviceable distribution map of these tools, I started to look further afield. It turned out that the average cleit in the Village Bay area was quite likely to contain a broken implement. Soon I noticed that broken stone implements also occurred in the walls of black-houses, stone clearance piles and the field walls and consumption dykes of the 19th century [ILLUSTRATION FOR FIGURES 1 & 2 OMITTED].
Building the cleits required a quantity of small stones, some used as chocks for the larger stones making up the walls. Among these small stones, I could usually find one or two broken implements, noticeable because of their colour. Most of the area within the 1830 head dyke is on acid granophyre - a stone which appears buff to white depending on the weather conditions. …