The third trench was very disappointing and we only investigated it in the 1996 season of work. We located it at the base of what we had believed to be the talus, but quickly discovered that bedrock lay almost directly beneath the turf. This was surprising, as Mesolithic material had been found eroding out of the cut face of the supposed talus around this point. We had assumed that most of the mound in front of the shelter consisted of occupation débris from the Mesolithic occupation of the site. Because of this, we had not expected the soil to be so shallow here; in fact, there was little more than turf over the top of the bedrock.
Survey work in 1997 (both geophysical and by test pitting) demonstrated that this part of the apparent débris mound was in fact a ridge of bedrock.