That ambiguity, between deliberate construction and slow collapse, is part of what makes this site quietly puzzling. One of the depressions along the exterior wall may, according to one reading of the evidence, be nothing more ancient than a sheep-shelter, a relatively recent hollow scraped out by farmers making use of whatever standing masonry remained.

The structures were recorded in detail by Judith Cuppage in 1986, who counted four or five of them in all. The largest sits against the inner face of the wall and measures roughly 2.25 metres by 1.6 metres internally, dropping to a depth of about 1.2 metres. A further D-shaped foundation abuts the outside of the wall just north of the southern entrance, with internal dimensions of around 2 metres by 2.75 metres and walls surviving to about 0.4 metres in height. Cuppage noted four or five structures but illustrated only four in her published plan. The fifth was identified through an earlier source: a plan published in 1899 by P.J. Lynch, which shows an additional structure to the south of the southern entrance, pressed against the outside of the promontory wall. Matching the structures visible on the ground today to the measurements and plans offered by Cuppage and Lynch is not entirely straightforward, and some discrepancy between the two accounts remains unresolved.