Over the last ten years, the institute has been excavating two extremely important sites: Serce Liman and Kas. Dating to approximately A.D. 1025, the shipwreck at Serce Liman has yielded the largest collection of Medieval Islamic glass yet known. The Bronze Age site at Kas, at present the oldest shipwreck in the world, dates to around 1350 B.C. and has produced a surprising quantity of glass. Today I will discuss the conservation methods and techniques used to preserve these finds.

The decomposition of glass is essentially a preferential leaching and diffusion of alkali ions, across a hydrated porous silica network Factors effecting this decomposition are:

1. The composition of the glass itself

2. The composition and pH of the attacking solution

3. The temperature of the attacking solution

4. Time.

The composition of the glass will determine its stability. Glass is composed of a network former -silica, and alkaline modifier such as sodium or potassium oxide, and calcium oxide, the positively charged ions of which will stabilize the negative ions of the alkali. The composition of most glass from archaeological contexts is unbalanced and much ancient glass is lost as a result of decomposition arising from excess alkali in the basic formula, insufficient calcium to stabilize it, or imperfect mixing of ingredients due to inadequate heating during manufacture. This results in alkali which is inadequately combined with the silica, and is thus susceptible to leaching out by the action of water. The remaining silica may be so minimal that chemical bonds between the glass components cannot be maintained and the glass will deteriorate.

With the exception of damage caused by dehydration, water, and above all accumulated moisture, is the most important factor initiating and sustaining forms of glass decay. The composition of the attacking solution, in this case a marine environment, is rarely stable and can alternately increase or limit amount of leaching taking place.

The pH of the attacking solution is important as glass retrieved from an acid environment often has an iridescent layer of leached silica. The alkali which diffuses out, is neutralized by the acid and few hydroxyl ions are available to react with the silica. This reaction leaves a thick layer; the alkali leaks out and the silica remains as a hydrated network, in short, water is holding the glass together.

Over time and under variable humidity many of these iridescent layers may build up into an 'Onion Skin' laminar formation and eventually exfoliate or fall off the glass.

Being extremely saline, sea water contains many chlorides, as well as oxygen and hydrogen sulphide readily generated through the activity of marine organisms and rotting vegetation.Glass from such an environment is less likely to have a weathered layer as the hydroxyl ions are available for reaction with the silica network and therefore do not cause a layering of hydrated silica. The reaction proceeds at a constant rate throughout the glass. The colouring matter in glass also becomes effected through the leaching of alkalis and a residue of undissolved metallic salts can be left on the surface. A partially devitrified glass surface may change the appearance of the vessel. Particularly striking effects of iridescent colour, may result from the defraction of light by the thin film or flakes which compose the surface of the vessel.

Moving water can cause a certain amount of erosion as can the build up of calcium carbonate resulting from decaying shell. This mixes with sand to form a rock hard coating which often cements a variety of objects or fragments together in a single mass. Much of the glass from Serce Liman was recovered concreted in such a way.

Before conservation or restoration takes place, thought must be given to what exactly is required of the conservator, and whether indeed the object merits complete restoration or just enough to enable photography or cataloguing. Each glass has unique qualities and different needs, it must therefore be examined in detail before conservation can proceed. There are a variety of conservation techniques, ranging from simple joining of fragments with adhesive, to drying, gap filling and supporting the vessels;

The Institute has excavated a wide range of glass from Serce Liman and Kas. From Kas, tiny glass beads and Bronze Age ingots, particularly interesting as they illustrate some of the earliest trading of raw glass. From Serce Liman, engraved glass beakers and bowls all have presented some interesting problems. In all cases glass excavated from a submerged site is kept wet until conservation can proceed. Immediate drying of such glass can cause irreparable damage due to devitrification.

All the fragmentary glass from Serce Liman were cleaned mechanically by hand, using brushes, scalpels and sometimes a hammer and chisel to separate the pieces and remove concretion and dirt. The fragments were then washed in changes of water to remove any chlorides. We start with tap water and progress to distilled water. Very little cleaning was needed, the glass was in extremely good condition with only a small amount of surface degradation. After thoroughly cleaning and washing, the glass was air dried. In most cases this proved to be a safe and efficient way of drying nearly all the glass.

Once dry, the glass was assembled, with joins held together with adhesive tape, the edges of the glass first being degreased with a solvent such as acetone. By using the tape, the correct shape and position of each piece was determined and accurately fitted. When the joins were all located, adhesive could be introduced into the cracks with a wooden stick.

The adhesive used for nearly all purposes is HXTAL NYL-1, an epoxy resin especially formulated for glass and ceramic conservation. Usually, the adhesive cures in about 4/6 days and the excess can be swabbed away with a solvent or picked off using a scalpel. Finally, the tape is removed.

Often simply drying and sticking the glass is not possible, perhaps the vessel is badly degraded or it may be held together only by sand and silt. In these cases great care must be taken.

The glass beads from Kas were found inside an amphora. One section became loose and it was decided to keep the mass wet and try and excavate some single beads. Desalinization was achieved through consecutive changes of water until a salt level comparable to that of distilled water was reached.

The salt level of the wash waters can be measured with the aid of a conductivity meter, an extremely useful gadget for any excavation. Distilled water is usually between 3-8 micro siemens. When a suitable level is reached and maintained for two readings, the water itself must be removed from the glass.

This is done by slowly replacing the water with acetone.

Initially 25% of the water is replaced with acetone, left for a day and then measured. One can safely use the conductivity meter with acetone as it does not attack the sensitive cell and the chloride content of the acetone itself can therefore be read. This is usually 0. Subsequently, 50% of the water is replaced and measured. This process continues until 100% acetone is achieved.

The object must now be consolidated. As the water holding the network together has been removed it must be replaced with something else. We use a consolidant called Paraloid B72 (Acryloid in America).

The object is either immersed, if this is possible, or brushed with the consolidant until saturation is achieved. The object is then removed and slowly air dried, usually under some loose polythene. The solvent will evaporate drawing the consolidant into the object itself.

These beads from Kas, some loose, some still in the concretion, have undergone treatment. Further excavation was not deemed possible.

Occasionally, after conservation, some restoration is required for aesthetic or structural reasons. This can be done in several ways. Two methods now used in Bodrum are:

1. A shallow mold, and where this is not possible, an inside former such as a balloon

In restoring a previously cleaned and assembled vessel which is in need of filling, a mold is made from a similar intact area.

This is done using warm dental wax or plasticine which is gently pressed against the vessel in order to obtain the required shape. The wax is then placed over the damaged area and necessary adjustments are made. It is extremely useful to outline the missing area on the wax with a felt pen, so filling with resin can be all the more accurate.

The wax is then allowed to harden and a plaster support is made to strengthen the mold. The mold is then coated with a thin layer of release agent, such as Vaseline or liquid soap. The mold is then filled with transparent or slightly tinted HYXTAL resin which has been allowed to cure slightly in order to facilitate control.

Using a small spatula and an overhead lamp as a heat source, the epoxy slowly dries to the required shape. While still flexible, the resin section is carefully removed from the wax, cleaned and fitted. It is important that the surface of the new piece is not too high, as sanding and polishing will reduce the thickness of the resin which can become too thin. The patch is secured using more of the same adhesive and surface irregularities are removed using grades of sand paper, finishing with an ultra fine sanding film. The new area is polished with a soft cloth or polishing buff on a flexible head with a little polishing paste.

An internal former is used only if the vessel cannot support itself and such restoration can be accomplished without distortion of the vessel. In many cases a balloon proves most useful. The balloon is inserted into the vessel, usually through the neck and gently inflated, being careful not to exert too much pressure on the fragile glass. Additional care must be taken to insure that the former is at the correct level for the edges and shape of the glass. The balloon is secured by knotting or tying the top and taping. The surface of the balloon is then coated with a release agent, Vaseline, making sure that every part

is covered and no Vaseline is adhering to the edges of the glass. Edges are then cleaned with a little acetone on a cotton wool swab.

Semi viscous HYXTAL resin is applied to the surface of the balloon. When the epoxy is completely cured the balloon is deflated and removed. Any fragments of balloon adhering to the inner surface can be removed gently and carefully with a scalpel and tweezers. The gap filled area is then sanded with graded sandpapers, and fine sanding film, and polished.

Sometimes gap filling by either of these methods is impractical. The object may be too large or of a complicated shape and cannot support itself.

After reassembling all available fragments, the bottle may still be unable to support its heavy weight. An inside former or support can be made of clear plexiglass. The plexiglass is bought in sheet form and readily available in various thicknesses.

An accurate template is made using cardboard, by taking measurements from the inside surface. Using a jigsaw, this template is then used as a guide in cutting out the actual plexiglass shape. One must be accurate and careful here as plexiglass is expensive and any marks or chips on the surface are difficult to remove.

Once the shape is cut, the edges can be trimmed using the flexible drive and a suitable burr head. Constant checking with the actual vessel is advised. If too little is removed from the edge of the plexiglass it can be rectified, but if too much is removed the piece is wasted. As one nears required size, graded sandpapers, or a sharp penknife may be used to refine the form to final shape. As the sand papers used become finer, the previously opaque edges become increasingly transparent and a final polish with a fine polishing paste is all that is needed.

The piece can now be placed inside the vessel, allowing accurate alignment of the fragments.