PREVIOUS   NEXT   CONTENTS   HOME 

1 Environmental and Research Background | 2 Post-medieval Buildings | 3 Earlier Vernacular Buildings | 4 Conclusions and Discussion

3.3.2 Wheelhouses

Plans of wheelhouses
Figure 40: Plans of wheelhouses in the Western isles: (a) Sollas, North Uist; (b) Clettraval, North Uist; (c) Baleshare, North Uist; (d) Tigh Talamhanta, Allasdale, Barra; (e) Cnip, Lewis; (f),(g) the Udal, North Uist (Parker Pearson and Sharples 1999, 6)

The domestic architectural form which preceded these cellular buildings in the Western Isles Middle Iron Age was the wheelhouse, so named because it displays a combination of circular ground plan and radially arranged piers that resemble the spokes of a wheel (Figs 40 and 41). They are found widely in the Western Isles, Shetland, Orkney (contra Armit 1996, 136) and the north coast of the mainland, and most examples were probably built between the late 1st century BC and the early 3rd century AD (Crawford 2002, 119). Compared with the architectural forms discussed above, considerably more published excavation data is available on wheelhouses and, consequently, it is less hypothetical to discuss the wheelhouse in general terms, although Crawford's (2002, 115) suggestion that we have uncovered 60% of the total number of wheelhouses ever built is, perhaps, erroneous (Crawford 2002, 115). In fact, given the regular movement of dune systems, the difficulty in dating surveyed settlement mounds, and the likelihood that continuation of settlement would make locating earlier sites difficult, it is likely that many more wheelhouse sites are still to be discovered. Currently, twenty-seven examples are known from the Western Isles, with only three of these on the long island of Lewis and Harris, at Cnip. The highest concentration of thirteen is in the Sollas parish in North Uist, with others in both South Uist and Barra (Fig. 42).

Plans of wheelhouses on South Uist
Figure 41: Plans of wheelhouses on South Uist: (a) Kilpheder; (b) A'Cheardach Bheag; (c) A'Cheardach Mhor; (d) Hornish Point; (e) Cill Donnain; (f) Glen Usinish (Parker Pearson and Sharples 1999, 5)

Distribution of wheelhouses in Western Isles
Figure 42: Distribution of wheelhouses in Western Isles. Key sites are labelled. Also see table 4.

There are two distinctive types of wheelhouse: those that are excavated into the sand, as with the cellular buildings discussed above, and those that are free-standing double-skinned buildings (Figs 43 and 44). Eighty-nine per cent of the examples from the Western Isles have free-standing radial piers which result in a circular passage or aisle between them and the inner wall (hence the alternative term 'aisled-roundhouse') (Crawford 2002). Many of these aisles have later been blocked and the examples in the Northern Isles tend to include piers that are bonded to the revetment wall in the primary build.

a free-standing wheelhouse
Figure 43: Author's reconstruction drawing of a free-standing wheelhouse. Based on Allasdale. Scale in metres

 Cnip wheelhouse
Figure 44: Author's reconstruction drawing of Cnip wheelhouse. Scale in metres

Other features include long, revetted entrances, while satellite cells, which are compared to side chapels by Crawford (2002, 121), are present at six sites in the Western Isles. Further evidence from Thomas' observations at Usinish (1868), supported by recent work at Cnip (Armit 1996, 139), suggests that the diagnostic radial piers supported corbelling over each cell, leaving only the central area to be roofed using a timber frame (Fig. 45). Nevertheless, in a recent synthesis, the excavator of Udal proposed that the central wheelhouse area may have been unroofed in its earliest phase (Crawford 2002, 123). Evidence for this assertion consisted of aeolianite deposits that thickly coated a line of 'flat 'stepping-stones' that front the bays' (ibid, 123). This has been interpreted as a drip course and the deposits are seen as the result of heavy, persistent dripping from the corbelled cell roofs. Of course, these deposits could have built up after the central roof had been removed, as in many cases the corbelling has lasted millennia while the central roof timbers, in particular, may have been recycled, as at Cnip (Armit 1996). This feature is perhaps more likely to be related to the demarcation of space, rather than the deflection of drips.

typical wheelhouse
Figure 45: Radial cells at Cnip during excavation (Harding 1996, 111). These are seen to exemplify a typical wheelhouse

Wheelhouses have commonly been thought of as domestic buildings; precursors to the cellular form, and descendants of the domestic broch tradition described below. Conversely, Crawford (2002, 127), relying largely on the common appearance of apparently votive deposits, the five examples of obelisks in the Western Isles, and the presence of cupboards (his 'aumbries'), coupled with the flimsy evidence for a lack of roofing, has suggested that wheelhouses may have been religious buildings, thereby introducing a dubious linguistic and functional relationship with later Christian buildings. Contra Crawford, I concur with Armit (1996) that the wheelhouse was principally domestic, as evidenced by the common appearance of hearths, material assemblages of mixed domestic type and their appearance at sites with long histories as domestic settlements.

The roofing is likely to have been a combination of corbelling over the cells, and a central, timber-framed conical roof, supporting a traditional Hebridean cover combining turf and roped thatch (Armit 1996, 139). Any turf and thatch covering would have to have continued over the corbelled area to protect the cells from damp, and, if McKenzie (1792, 288) is to be believed, the cell roofs may have been coated in clay. In those examples existing as free-standing buildings, this arrangement would result in a relatively flat area remarkably similar to the exposed wall-top or tobhta of post-medieval blackhouses, allowing easy maintenance of the conical thatched roof. Combining a central timber roof with stone corbelling meant a significantly reduced need for timber. For example, without piers, the area of 29 square metres at Alt Chrisal in Barra would have required a 45° conical roof with a surface area of 41 square metres with 4.2m rafters. Using the design described above, Alt Chrisal may have had a conical roof with a surface area of only 15.1 square metres and 2.6m rafters, the design therefore significantly affecting the amount of timber required.

The size of wheelhouses in the Western Isles varies greatly, with internal areas of less than 30 square metres to over 100 square metres (see Table 4). Even at their smallest, they are comparable with typical domestic spaces in the post-medieval period. Unfortunately, evidence for the use of this space within the wheelhouse is limited. Examples such as Allasdale, Barra (Young 1953), and Clettraval, North Uist (Scott 1948), are associated with external structures that may be contemporary, suggesting perhaps that much of the activity associated with the post-medieval barn and byre may have taken place outside the wheelhouse. However, there have been suggestions that particular cells may be associated with specific tasks and it is also possible to see a bipartite division of space within the wheelhouse itself; in Cunliffe's words:

'Two areas were created: a central area with a high roof line where the principal activities of the house were carried out, and a surrounding area between the posts [piers] and the wall of the hut which would have served as storage or a sleeping place.' (1978, 175)

The use of space within wheelhouses is further discussed in the section on brochs.


 PREVIOUS   NEXT   CONTENTS   HOME 

© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions | File last updated: Tues Feb 28 2006