Investigations in wood identification

My involvement with the Samarra project began in 2013 when I was asked to help identify the timber of some wooden archaeological elements from 9th-century Iraq. The artifacts were a combination of mostly flat planks and a few thicker, carved wooden elements. All have remnants of paint on the show surface. While historical sources say that these wooden elements are carved from imported teak, to our knowledge this had never been tested or verified scientifically. The occasion of the V&A's Samarra project seemed like the right time to pursue the most comprehensive type of wood identification analysis. We therefore contacted Dr Peter Gasson at the Royal Botanical Gardens in Kew.  Peter is an expert on microscopic wood identification and had previously taught me the basic techniques. Due to his extensive experience and Kew's comprehensive reference slide collection, he seemed like the best person for us to work with. Happily, he confirmed that he could analyse the samples for us.

Wood identification is done by comparing defined features of the timber with known samples. While many features can be distinguished by eye or with the help of a hand loop, the most comprehensive identification is done microscopically. While this method gives the most reliable results, it is not non-destructive and requires a sample of material to be taken from the object. The ideal sample size is about 1 cm cube.  On many museum objects, a sample this size can be difficult to justify. In this case, it was decided that the potential benefit to knowledge was worth the sacrifice of the material. We decided to take samples from the back surfaces of two representative pieces, keeping in mind that if they were ever to be placed on public view in the future, the sample sites would not be visible; potentially they could even be plugged if that was considered desirable.

The V&A's collection of wooden fragments from Samarra can roughly be divided into two types: flat, relatively thin, painted boards and thick, carved fragments with minimal paint left on the surface. One of each type was chosen for sampling based on their representative qualities of the group as a whole, and the fact they had been carved and decorated in different techniques from each other. 

A.129-1922. Photo © Wei Kao

A flat, painted piece (A.129-1922) was chosen because it had already suffered some splitting and wood loss. It also had enough thickness to accommodate the sample size required. We also selected a carved, three dimensional piece (A.134-1922) as an example of a different production technique, and because its flat reverse presented a large surface area for sampling.

A.134-1922. Photo © Wei Kao

Two protocols were tested for removing the samples. The first was to sample from the edge of the object, using a combination of hand-held drill, saw, and chisel. The second was to use a wood-boring bit in a pillar drill. The latter seemed most invasive due to the wastage caused by the drill bit. We began by testing the sampling method on a piece of ash that was not a Museum object. Both tests worked well.  Due to the different dimensions of our sample pieces we decided to use both methods; the edge sampling was chosen for the painted plank, and drilling chosen for the thick, carved piece.  

With the help of two interns, Wei Kao from the Division of Wooden Objects and Oil Painting Conservation, Institute of Conservation of Cultural Relics and Museology, Tainan National University of the Arts (TNNUA), Taiwan and Hsun-Hui Hsu, MSc student in Conservation for Archaeology and Museums, UCL, the two wooden objects were prepared for sampling. Here we are looking a bit nervous, about to take the samples from our two objects, with our successful test piece visible on the workbench at the right:
 

Pictured: Dana Melchar (left), Hsin-Hui Hsu (centre), Wei Kao (right). Photo © Mariam Rosser-Owen

We began with A.129-1922 by pencilling in the sample area on back of the plank. 

Photo © Mariam Rosser-Owen

With the fragment clamped firmly in place on a flat surface, I began by drilling the perimeter of our sample.  

Photo © Wei Kao

Next, I used a small saw to cut the side edges. 

Photo © Wei Kao

Then a small chisel and mallet to remove the sample, tapping as gently as possible!

Photo © Mariam Rosser-Owen

Here's what the object looked like after removing our sample.

Photo © Mariam Rosser-Owen

Unfortunately, our sample broke into four pieces as we removed it from the object: one large and three small pieces.  Luckily, our large section was still big enough for Peter Gasson to use for microscopic identification.

Photo © Mariam Rosser-Owen

We then moved on to sampling the thicker, carved fragment.  We clamped the fragment to the table of our pillar drill, set the wood boring bit to the correct depth, and then turned the machine on and made the hole.
 

Photo © Mariam Rosser-Owen

After the drilling:
 

Photo © Mariam Rosser-Owen

With the sample removed:

Photo © Mariam Rosser-Owen

It was a relief when the sampling was completed and we had two great samples to send to Kew! I was able to hand deliver the samples personally and fill Peter in on the research we were conducting. 

We heard the results from Peter a few weeks after he had received the samples. Based on the features he observed, the clear result was teak: specifically, the Tectona species, which is native to South and Southeast Asia (it grows in Central South Asia, India, Pakistan, Sri Lanka, Burma, Southeast Asia and the Pacific, Thailand, Laos, Vietnam and Cambodia: see here for more information). This confirms the information provided in the historical sources and makes sense in terms of trade routes at the time.

Peter noted that the sample from A.129-1922 was very typical in appearance for teak. It showed narrow growth rings, indicating that the tree was grown in an area or time period without much annual rainfall. This results in a denser wood, with a less porous surface. This makes sense of the fact that this wood was selected for painting by the Samarra woodworkers. As we learned through the pigment analysis that we also conducted - together with our V&A colleague Dr Lucia Burgio - no base or primer layer was used for the painting, and the pigments were painted directly onto the surface of the wood.

On the other hand, the sample from the carved object  (A.134-1922) appeared more like intentionally cultivated wood. It exhibited more widely spaced growth rings, indicating that water was more readily available wherever this tree grew, and that it may have come from a faster-growing tree, possibly a younger tree that was cut down before it matured. This results in a less dense wood, which would be easier to carve. 

It is intriguing that the Samarra woodworkers seem to have been experienced enough in handling teak to know its characteristics, especially its density, and to have selected exactly which timber would be better suited for painting and which for carving. It gives us another insight into the craftsmen who built the palace-city.

With all best wishes to the Samarra blog readers for the holidays! We look forward to sharing more insights and discoveries here in 2015!

Dana Melchar
Senior Furniture Conservator, V&A