Wood can be preserved for long periods of time under anoxic, waterlogged conditions. Whilst the absence of oxygen reduces the rate of decomposition, wood is always modified in the burial environment. Prior to the oxygen being used up, fungal and microbial action results in the loss of cellulose, an important structural polymer in the wood. As a result the wood is very much weakened -under extreme conditions, only a lignin skeleton is left - and whilst this may retain all of the detail of the original wooden artefact, its strength is less than 1% of that of modern wood.

The conservation problem is to dry the wood without it being destroyed by this process.

Supercritical drying, a new technique developed at St Andrews avoids the shrinkage problem by replacing the liquid in the wood with a high-density gas or supercritical fluid. The supercritical fluid can be removed from the wood by decompression without forming a liquid phase. As shrinkage is due to surface tension forces at a liquid surface, supercritical drying does not damage the artefact.

The supercritical fluid we use is carbon dioxide, which is cheap and safe, but will not mix with water. In consequence the water in the artefact must first be replaced with an organic solvent - methanol. At high pressure the carbon dioxide's density is similar to that of common liquids, and it readily dissolves the methanol.

Below are some of the samples conserved by supercritical drying, which demonstrate some of the advantages of this technique.

The technique of supercritical drying, allows wood in any condition to be dried without modifying the methodology. So that heavily degraded material, with no inter-fibre adhesion, (Mesolithic timber above), and sound timber, such as that from the Dartmouth (see examples), both responded well. Other methods rely on replacing water within the wood structure with a polymer, which cannot normally be introduced into cork, and some other dense woods, which remain impermeable. Further, polymers and other foreign material accelerate the corrosion of metal, by dissolving the protective surface oxide, or simply by keeping it wet.

Figure 5. A cork bung from a 'dog buoy' (an inflated dog skin, commonly used to lift fishing gear in the past). Successfully conserved by supercritical drying after attempts at freeze drying had failed.	Figure 6. Compass base from the Swan. Composed of a heavily degraded turned wood platen, with a lead righting weight. A copper pin projects from the centre of the lead weight, this acted as the bearing on which the compass card would have turned.

Under some circumstances supercritical drying offers advantages over other methods. Firstly in helping to stabilise the material, for example, bone. In this instance the carbon dioxide used in the technique converts any organic calcium salts into calcium carbonate. If left in place, these salts can result in the formation of 'Bynes disease', a white efflorescence which can result in loss of surface detail. Secondly, the absence of any foreign material means that the internal structure of the wood is not obscured.

Figure 7. Human vertebra from the Swan, after conservation by supercritical drying.	Figure 8. Cross section of an ash spear shaft (Fraxinus excelsior) from Nydam Mose, Denmark.

Tool Marks As the true potential of wetland archaeology is becoming recognised there is increasing interest in evaluating the tool marks left by the makers (Figure 9, Figure 10). If two timbers bear the same tool marks, then they are likely to have been part of the same construction. This can be a very useful aid to understanding crannog construction. The sites of these lake dwellings have normally been re-used many times over periods of several thousand years. This leaves a confusing jumble of timber on the lake bottom, and any technique, which helps to relate this material, is potentially helpful. Tool mark analysis has not yet fulfilled all of its promise, but is increasingly finding utility as a check on dendro-chronology and other analytical techniques.

Figure 9. Marks left by notches of an axe head

Figure 10. Sketch of the house post from above, showing two facets (each produced by a single axe blow), and the signature, an array of parallel marks in the wood produced by nicks and burrs on the axe blade. The tree rings cut through by the axe are shown by the dotted lines.

Below, two early worked timbers showing tool marks, conserved by supercritical drying.

Figure 11. Tool marked house post, photo in strong cross illumination	Figure 12. Early bronze age coppice wedge from the Seaton 3 excavation.