Carnbaan

NR86NE 38 853 682

On a rocky outcrop are the foundations of a small rectangular building, possibly the peel built at West Tarbert on the orders of Robert the Bruce in 1326 (see NR86NE 21). Tarbert means in Gaelic an 'over-bringing', a place where boats were hauled overland from one body of water to another. The portage across the narrow, low-lying isthmus at tarbert was the scene of an incident recorded in the Orkneyinga Saga: in 1097 Magnus Barelegs was drawn across it in a skiff in order to claim Kintyre for Norway. The saga writer commented that ships were regularly hauled across at this place.

Traditionally, ships were hauled across using tree-trunks as rollers, so when deep ploughing of the field brought up several large pieces of timber about 1990, it was conjectured that these might be the remnants of the original surface of the portage, laid down where the surface was boggy. One of these timbers remained in situ, some 55cm below the present ground surface, in the face of a drainage ditch cut through the centre of the field. The ditch had also cut through a discrete area of large boulders and stones. It was conjectured that these stones might be a surface relating to the use of the building on the outcrop. Some six timbers were sampled for dendrochronological analysis and the section in the drainage ditch was cleaned and recorded.

The timbers were all oak (Quercus sp.) and all were decayed to varying degree due to exposure and desiccation. As a result none retained any sapwood and the heartwood/sapwood boundary was also damaged and decayed. The timbers were long and very straight-grained, the longest being 4.96m in length. There were only two branch scars visible along this length. Only forest-grown oaks would produce straight-grained boles of this length with few branch attachments. Only fragments of the original cross-section of each timber remained. As there was no toolworking visible, the timbers had either been shaped by cleaving or had spilt naturally.

Cross-sections of each timber, 10cm wide, were sampled and the surfaces were sampled and the surfaces were prepared so that the tree-ring pattern was clear. This was done using a razor blade to produce a smooth glossy surface into which powdered chalk was rubbed to highlight the large springwood pores. The samples were then measured using standard dendrochronological equipment and cross-matched. The six samples produced promising long ring-patterns and pairs of smaples cross-matched well together: Samples 1 and 2 correlated well, but the correlation between Samples 5 and 6 was so close as to suggest an origin in the same tree. The samples in each pair were averaged together to form two sub-masters and were then run against each other and the remaining Samples 3 and 4. No further matches were found. The four sequences, Samples 3 and 4 and sub-masters 1/2 and 5/6 were then run against a suite of calendrically-dated chronologies from Ireland, Scotland and northern England spanning the last two millenia. However, no significant and consistent position of match could be found, and so the Tarbert timbers must remain undated.

The southern face of the drainage ditch was cleaned and recorded along a 20m section. The earliest feature visible is a pile of large boulders and stones, 7m across and 0.75m high, which had been dumped directly on to a subsoil of raised beach sands and gravels. To the W this feature peters out and the overlying ploughsoil lies directly on the subsoil. To the E a series of deposits had built up behind the pile of stones. A peat with bands of grey clay had developed over the subsoil. The oaken timber, Sample 4, lay on one of the grey clay laters.

The field is poorly drained and the bands of grey clay probably represent episodic flooding. The stone pile may represent an attempt to provide a stable surface over the boggy ground; if it is a surface it is running roughly N-S in a line from the rocky outcrop to the higher land along which the present A83 runs. Peat started developing on the poorly-drained land and the timbers may have been a further attempt to provide a dry surface. Unfortunately, in the absence of dendrochronological dates these features cannot be dated.

B A Crone 1994; AOC (Scotland) Ltd 1995.