During excavation of a trench known as Cutting 1, archaeologists uncovered a cluster of three possible structures in an unusually exposed stretch of ground. One of them, recorded as Structure 4, sits just south of a companion feature and presents a puzzle that even its excavators could not fully resolve: the outline of a building reduced to something barely recognisable as one.

What remained was a roughly circular spread of flagstone paving, measuring around 4.7 metres east to west and 4.4 metres north to south, enclosed on three sides by a vague arc of collapsed rubble walling. The northern stretch was the clearest, where stones still formed a low bank, and a few overlapping pieces hinted at a coursed foundation wall that had either fallen or been deliberately levelled at some point. Along the south-west, rubble had tumbled inward over the paving itself. The floor had more than one story to tell: in places, two layers of flagstones were laid one over the other, probably to even out an uneven surface, and in the northern part of the floor, earlier flags had been lifted and replaced. A thin layer of compacted clay, flecked with charcoal, marked the boundary between the two floor phases. The finds were sparse but suggestive; two fragments of clay moulds, the kind used in prehistoric metalworking to cast objects in bronze or other metals, were recovered from the floor layer and the rubble walling respectively. The site is considered prehistoric in date, though the excavator, reporting in Cotter 2012, was careful to note that Structure 4 and its two neighbours need not all belong to the Late Bronze Age specifically.