A repeat analysis of 21 sites in the Mediterranean, brought about by the need for obtaining samples for dendro-dating also looked at the state of preservation of the wrecks some 20 years after they had been originally surveyed. This survey found that the reburial technique had only a 33% success rate. Shallow sites with low sediment deposition were worst - on one site only one timber remained of the original wreck. A major cause of damage was though to be diver activity.

Deep wreck sites (>30m) with high sedimentation rates (for example in estuaries) were best preserved. Artificial stabilisation with sandbags was found to be only mediocre, due to water pockets trapped beneath the bags allowing burrowing animals access.

Using the Lyneas site (pronounced 'Lunas') in Denmark as a location for experimental models for wood deterioration during reburial. Samples included oak test boards and standard cellulose cloth. These were mounted on pine poles above the sediment, just covered in sediment and buried at 50 cm. The tests were designed to coincide with peaks in microbial activity, which occur in April and December in temperate waters. Monitored parameters include nitrogen as total [N] and as species - NH4, NO3 and NO2. Nitrogen is often considered to be the limiting nutrient in wood, with ammonia being the end product in reducing environments, and nitrite and nitrate in the presence of oxygen.

Results showed that the pine poles were destroyed above the sediment surface within 16 weeks by Teredo sp. (possibly Teredo navalis), after 32 weeks exposed oak timber was also destroyed. Immediately below the surface some activity of tunnelling bacteria was noticed, but no soft rot or shipworm. The buried timber showed a loss of 20% of their strength over 64 weeks, irrespective of burial depth.

Analysis indicated that nitrate was still present in the sediment at a depth of 50 cm after 64 weeks, indicating these sediments were still oxygenated. The C/N ratio was put forward as a possible monitor for the biological activity of the sediment relevant to the preservation of archaeological material which might be reburied within it. Problems were encountered in measuring redox potential of sediments.

Use of silver/copper electrode ionisation to clean water used to preserve the site of a medieval mint. Water treated with the ionisation technique was misted onto the site at regular intervals to keep the site damp. The site is enclosed beneath a modern building and open to the public. No further conservation was attempted. No prediction for the long term preservation of the site was given, but as the alternative was allowing the site to dry out, it was thought that this experimental technique was preferable. Modern wood blocks were installed to monitor for Cu and Ag uptake, and attack by fungi and other bugs.

The Zakynthos wreck being excavated by MARE and the Greek authorities since 1984 was the site of trials using geotextiles for in situ protection of material in the excavation off-season. The reef about the wreck maintained a stable sedimentary environment, and sandbagging was thought to allow continued access and development of marine borers, which gave the most rapid deterioration of the site. The organisms of interest were Teredo, and arthropods such as chelura (shallow tunnels) and limnoria (surface indentations), the arthropods are always found together.

The intention was to provide a physical barrier to prevent access by the organisms, without totally preventing water movement, and polypropelene/polyethylene geotextiles were chosen in consequence. These offer the advantages of being cheap, easily applied (slightly positively buoyant, so weighed down with rocks) and having a low environmental impact. By contrast biocides have been known to cause skin irritation in divers.

In experimental studies the Geotextiles from Exxon tested included TERRAM 500s, TERRAM 1000 through 4000 (reducing pore sizes). In addition T1-19V, a sealed aluminised plastic wrapping, and T-44A, a wide mesh material. Bags were double stitched or heat sealed (T1-19V) about small wooden test planks and mounted exposed to seawater on racks at several locations about the site. The racks were sampled at intervals of 6 months, 1 year and 2 years, after which time the control (unwrapped) had completely disintegrated. In many instances the wide mesh bag was worse than the control, as this provided protection for the borers, but not for the wood. Terram 2000 and 4000 were found to provide good protection from borer attack.

Over a period of one year a natural sediment cover was established over the textile, which blended into the surroundings. Redox potential measurements through the textile indicated that anoxic conditions were re-established within a year of its application. Note the textiles are not effective in preventing fungal or bacterial activity.

Under favourable conditions wood can be preserved in situ for 80M years in an un-petrified state. Despite years of effort, no bacteria responsible for degredation of wood have been isolated in pure culture. Organisms which attack wood include:

1. White Rot - some types are lignin specific.

2. Brown Rot - observed by the loss in birefringence in polarised light in the S1 and S2 cell wall layers. This type is not aquatic, but may have occurred prior to immersion.

3. Soft Rot - brown colour, actually timber is very hard in the dry state, but spongy when wet. This attack leads to the appearance of chains of cavities within the S1 and S2 layers of the cell wall.

The above fungi cannot survive in anaerobic conditions, but pre-digestion by them may aid further attack by bacteria.

1. Tunnelling - require more oxygen than erosion bacteria, and are very common in high compost content soils. Their activity is usually masked by the presence of fungi.

2. Erosion - can operate at very low oxygen poartial pressures, but not in the total absence of oxygen. This type is probably the most important factor leading to the decay of waterlogged archaeological wood in 'stable' environments.

Sulphate reducer bacteria are not thought to have a direct effect on wood, but side products of their activity, notably changes in pH, may result in damage.

Laboratory experiments on the deterioration of modern wood in oxygenated sand and soil. Pine test samples half buried in damp, but not saturated, sand or soil. 1/3 samples covered completely with the same material, 1/3 samples covered with geotextile and more soil or sand and 1/3 covered with sawdust and capped with clay. Sawdust was intended to be a good environment for cellulitic bugs and fungi, activity there was expected to absorb nutrients that would be required for the attack on wood.

Results indicated that un-sterilised top soil was 6x more aggressive than sand or clay. All samples lost weight over time and that geotextile and sawdust were about equivalent as protective mechanisms.