A Glass Puzzle from Tell Fara

Emily Williams

Image Of The Glass Sherds Before Treatment. Copyright UCL Institute of Archaeology Collections

When I first came across the box full of archaeological glass fragments in the conservation lab, it raised a strong sense of curiosity and concern – I did not want the box of bits to go back to the collections store in the same condition, and without knowing what this golden iridescent glass used to be. The box contained 19 fragments, too few to make up a whole vessel, so early on, I had to abandon any idea of being able to reconstruct a complete object. After many hours of research, and what felt like the longest, most complicated puzzle I had ever done, I was however able to piece enough fragments together to infer the shape of a bottle.

Reconstructing the bottle. Copyright UCL Institute of Archaeology Collections

Once I had identified that the fragments formed a bottle, I attempted to reconstruct it. The glass was however too fragile to handle without the surface layers becoming detached. Before I could do anything else, I needed to consolidate the surface. This was done using very low percentage of Paraloid B 72 in acetone. I carefully applied this solution so that the polymer went under the delaminating flakes in order to reattach them to the underlying glass.

Making the Fill using Japanese Tissue. Copyright UCL Institute of Archaeology Collections

After I consolidated the surface, I put the fragments together with a more concentrated solution of Paraloid B 72. Stronger epoxy adhesives are often used to reconstruct glass objects in collections (like Hxtal NYL-1), but was not used with this archaeological object as the treatment would not have been so easily reversible. The potential for conservators in the future to retreat our conserved objects is always considered when we make decisions about our treatments now. I then used watercolour paints to colour Japanese tissue a very pale blue/green and sandy yellow in order to match the colour of the glass underneath the layers of corrosion. The tissues were shaped by hand, cut to size, and then attached as a support to the reassembled glass fragments.

The Reconstructed Base of the Bottle. Copyright UCL Institute of Archaeology Collections

For the upper section of the bottle, I experimented using a light bulb as a base to create my rounded fill. Shaping Japanese tissue so that it remains curved can be quite tricky. My lab partner Robert suggested I coat the light bulb in a thin layer of silicone to create a barrier between the two, this allowed me to remove the tissue from the light bulb surface once it had dried. I used T40 Silicone Rubber from Tiranti, which is a two part thixotropic mixture with a curing time of approximately 30 minutes.


Using the Silicone covered light bulb as a mould for a Fill

One complicated aspect of the reconstruction was having to figure out how to reconstruct a closed vessel when the only opening was through the very small and narrow neck of the bottle. Once I had all my tissue supports in place, I inserted a bent wooden stick with a flattened silicone tip so I could reactivate the adhesive with acetone and use the silicone tip to press the tissue onto the glass from inside the bottle.

Glass Irridescence On The Inside Of The Bottle. Copyright UCL Institute of Archaeology Collections

Finally, the glass bottle reconstruction is now complete! The object will be returned to the UCL Institute of Archaeology’s Palestine-Petrie Collection, and is now readily available for teaching and research.

Image of the Bottle After Treatment. Copyright UCL Institute of Archaeology Collections


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s