By Beatrice Demarchi, Research Fellow at the Department of Archaeology, and Dr Julie Wilson, Lecturer at the Department of Chemistry. University of York.
This article is part of our series: a day in the software life, in which we ask researchers from all disciplines to discuss the tools that make their research possible.
Interdisciplinary research between fields as diverse as biochemistry, physics and archaeology can help us decipher the most amazing things, for example the choices that our ancestors made thousands of years ago when deciding how to adorn their dead.
People have been using personal ornaments (jewellery) for tens of thousands of years to demonstrate their status in a group or society. If we are lucky, we may find these ornaments in graves or caves and can speculate as to their meaning and provenance. Were these “jewels” made with local raw materials? Or were they traded in exchange for exotic goods? Each of these possibilities tells us something about the environment that people lived in and how they chose to exploit it. However it is often very difficult to identify the raw material when the ornaments are both sculpted and also found in a degraded state.
Such a mystery presented itself to Suffolk County Council’s Finds and Post-Excavation Manager, Richenda Goffin and the archaeologists who excavated the burial of a young woman at the site of Great Cornard. Lying in the ground next to the woman’s head were thousands of black and white beads, placed in a zebra-like pattern, and some amber beads. Clearly this had been a very complex creation - maybe a necklace or chest ornament, a question Alison Sheridan at the National Museum of Scotland is currently trying to solve by reconstructing the ornament. The black beads were soon identified as Whitby jet, but the white beads were less obvious. Sonia O’Connor, based at the University of Bradford, used electron microscopy to examine the microstructure and found that the white beads were made from mollusc shell.
However, we wanted to know the type of mollusc. If these beads were fashioned from some exotic Mediterranean shell, such as the pink thorny oyster (known to have a special symbolic significance in prehistory), then this would mean direct contact and trade between the Great Cornard people and the Mediterranean world. This was not impossible in 2200 BC and the sourcing of exotic shells might also have been appropriate for what appears to be a very special ornament.
In order to determine the type of shell used, scientists from the Departments of Chemistry, Physics, Archaeology and Mathematics worked together. Kirsty Penkman and I [Beatrice Demarchi] had been doing research on mollusc shells for quite a while at the NERC-recognised facilities of the NEaar laboratory, and we know that shells trap a tiny fraction of proteins (responsible for their biomineralisation) within the calcium carbonate crystals and that this intra-crystalline fraction is stable.
In fact, we use the extent of degradation of these proteins for dating purposes. We had noticed that the amino acid composition of this fraction seemed to vary with the taxon of the molluscs. Julie analysed the data on the amino acid composition of nearly 800 shell samples of known species. We discovered that, although the classification is not perfect due to natural variability, we could still identify the genus of a shell correctly 77% of the time. For the classification, Julie used statistical pattern recognition algorithms, such as Learning Vector Quantisation (LVQ) available as a package for academic use from the LVQ Programming Team of the Helsinki University of Technology.
The shell data were separated into two sets, a training set on which the algorithm is trained to recognise the patterns associated with a particular class, and a test set, not used during training, in order to test the algorithm. In the LVQ algorithm, the training data is used to obtain a set of vectors, containing information on the relative amounts of amino acids, representing each genus being considered. The test data were classified by comparison with this representative set of vectors. As well as being able to identify the genus of the samples in the test set correctly most of the time, we could see, by considering different taxonomic levels, that closely related taxa (but not the same genus) had similar patterns.
Once we knew that we could classify shells reliably, the trained LVQ algorithm was used to classify six shell beads from the Great Cornard ornament. We discovered that 5 were very similar to each other in amino acid composition, and the sixth was very different. At least two different shell taxa had been used. We found a very good match with the common dogwhelk (Nucella lapillus) for the first 5 beads, and with the tusk shell (Antalis genus) for the sixth bead. Our results from the biomolecular analysis were supported by data from other techniques. Yvette Hancock, alongside Raquel and Andre Ponzoni, used Raman spectroscopy to show that the mineralogy of the beads matched that of the proposed mollusc shells. Also, Sonia O’Connor compared the microstructure of beads to that of the candidate shells, using both light microscopy and Scanning Electron Microscopy, and confirmed that dogwhelk and tusk shell could indeed be the genera used to create the beads.
Although we certainly cannot be 100% sure of our identifications (for example, instead of dogwhelk we could be looking at another closely related species that we have not included in our database), we think it is reasonable to hypothesise that the people of Great Cornard did not travel very far to source the raw material for making this precious necklace. The UK’s shores provide abundant specimens of dogwhelk, and tusk shells are not unknown finds either. As the photograph shows, their hollow shape is perfect for making beads – and indeed the high frequency of tusk shell beads found in prehistoric sites shows that our ancestors had realised this a long time ago.
So what is the conclusion? Well, if you‘ve just tried to save some money on a gift for your girlfriend by buying cheap and local jewellery instead of exotic – think again. The archaeologists will get you - even if it takes 4000 years!