Chasing the Taxidermist’s Tail: Using chemistry to reconstruct specimen collection data in natural history museums

Celia Cramer: I am a PhD candidate at the School of Chemistry in Sydney University. And I'm talking to you today from Canberra. So I thought I'd like to start by paying respects to the traditional owners of the land I'm on and to the Ngunnawal elders past, present and emerging.

So today, I thought I might speak to you a little bit about my ongoing research in which I asked the question; Is it possible to find the field collector of a natural history specimen through the spectroscopic analysis of its preservative residues? So I'll give you a bit of background at first on natural history collections, on historic preservatives and on spectroscopic analysis before I share with you some of the cool things that I've learned so far.

So, as I'm sure many of you know, natural history collections in museums and universities are a storehouse of ecological data that offer unique research opportunities, these massive collections abroad in diversity, their wide and geographic scope, and extend over a really long periods of time. And it's this temporal aspect that makes museum collections so beneficial to studies that change over time. And these resources are being well used in 2015, Brian Maclean and his team found that museum specimens were utilised in 25% of papers published from the Journal of mammalogy in 2005 to 2015. It's estimated there are about 3 billion natural history specimens in the world's museums and university collections, which means it's roughly the number of natural history specimens is roughly equal to all the other objects in all the other galleries and museums put together, which is a remarkably large number.

But for research, for education, and even for display, the value of each specimen is enhanced, or limited by the quality of its contextual data. And this data typically includes things like an animal's the animals location, bit of information on its habitat, the date of location, the field collectors name, and any notes on research or scientific testing that's been done to the specimen since it joined the museum collection. Museums with unreliable, inconsistent or even missing data often have to be excluded from research regardless of whether or not they're really rare or really unique.

Numerous publications have noted the difficulty of missing data when they're working with museum specimens, the US in the US. One study found that 9% of mammal specimens across the American continents had no year of collection, and Malaney & Cook found that 3% of mammal specimens in US collections were missing their location data. And if we extrapolate this last percentage into a global context, that means that there are about 90 million specimens that must be excluded from research just due to missing location data. While no studies have ever looked into the missing data and Australian collections, a viewing of Atlas of Living Australia shows us that 25% of Australian mammals and 9% of birds have no year of collection, while 72% of Australian mammals and 59% of birds have no data for field collector. And it appears that the oldest specimens are more likely to have been separated from their data than the younger ones. And it's these older specimens that really can offer the most in research where time is a factor. So it's this very issue that gave rise to the ARC link project that I'm a part of the project aims to match 18th and 19th century field collectors with trade records, and Expedition notes to determine some of that contextual data like date and location. It's a large multidisciplinary team with a number of different approaches being used to resolve the problem and explore all sorts of aspects about them. My role is to investigate the possibility of using the evidence within the specimen itself, the preservatives to identify the field collector.

So I'll start with a little bit about the preservation of skin. Skin rapidly decays without some form of preservation or preservative. The cells aren't enzymes and bacteria typically begin the decomposition process. The period of time it takes for skin to decompose depends on the surrounding temperature and humidity, as well as its proximity to insects and other animals. But to try and put a timeframe around it; hair can begin to fall from the skin within a couple of hours in warm conditions. And a whole body can take up to a year to decay in cooler conditions. So skin can be preserved by creating a local environment that's inhospitable to the growth of micro organisms, and making the skin unpalatable and excluding animals and insects as much as possible. And when it's well done, skin can remain intact for centuries. This electron microscope image that you can see, even after 130 years, there's little bundles of collagen here and here. And collagen is a structural protein found in skin and you can see these areas where the environment has been restricted, so it hasn't had too much exposure to oxygen, which is very, very cool.

So For the 17th and 18th centuries, European natural history field collectors realised that a single preservative alone just wasn't going to do it, it wasn't sufficient to ensure the preservation of skin. So they started combining ingredients and developed recipes. These preservative recipes were initially kept as a secret shared only between colleagues and letters. But by the mid 18th century, the sharing of these instructions became an industry in its own right, helped along by the repeal of taxes on paper and publications. So over 250 instruction manuals were published in the 18th and 19th century on taxidermy. Free or subsidised pamphlets were also published from enthusiast clubs and museums, and there was republication of some of those recipes in newspapers. And this all allowed the general public to help participate in national history collecting as well. In recent decades four authors mostly in the USA, have tried to catalog all the diverse ingredients that are in preservative recipes.

So the methods of applying preservatives were varied, and they varied on cultural group and also over the period of time. But this flowchart shows some of the most common methods used and documented during the 19th century, when Australian specimen collecting was in its heyday. So after flaying and cleaning, the skin was either dry salted or put in a pickle bath, it was left to dry and then a preservative mixture was applied. At this point in time, the skin would be arranged sometimes left foot over right to show gender and then stuffed to give it a kind of shape and fill out the flesh or the skin and then labeled. And later sometime later, sometimes many hundreds of years later, the skin will be rehydrated and mounted as taxidermy or it'll be just stored as a steady skin so people could research it. So when it comes to looking back from the 21st century to identify field collectors of the past, the method up to now has really been focused on stylistic analysis, similar to the method used to identify artworks by handling of paint or documents by handwriting. So the method seeks to match known taxidermy samples with unknown ones using external appearance, and also the internal wiring like in this X-ray image. And it's all based on the assumption that an individual's art or craft practice as a taxidermist changes little over time. Chemical analysis of Natural History specimens has been undertaken in recent decades, but the focus really has been on identifying and quantifying the poisons that were mostly applied as pesticides rather than the preservatives themselves. But encouragingly, spectroscopic analysis of mummies and parchments in historical others, have shown that a number of preservatives can be detected and identified. However, many of the successful techniques have required physical sampling. And that has had associated damage with the artifact itself. So it's not ideal.

My research aims to build on all of these old studies to develop a method of analysis that doesn't require physical sampling, but uses non-invasive and non-destructive spectroscopic techniques to support the goal of in perpetuity, preservation and museum objects. It's a bit like a classic whodunit, but with the analytical techniques normally reserved for forensic analysis of crime scenes. But as you would expect, it's not without its complexities. Because the chemical element evidence within a natural history specimen is a mixture of the biological material, the environmental contaminants acquired during its life, preservatives applied at collection, and the contaminants and pesticides acquired during transport, trade, and even in the museum environment. So for example, if we consider the typical application of a preservative at collection, it's applied normally by by a brush or by rubbing the preservative in with your hands. But there will naturally be a variation in concentration where your hand or your brush is fully loaded with preservatives when compared to when it's nearly empty. In addition to this, the natural variation in the porosity of skin means that in some areas, preservatives will penetrate more thoroughly than in others. So all these factors together result in a remarkable amount of variation across a single specimen. And I will need to unravel the mixture before you can kind of interpret the chemical evidence.

This point and shoot method that's recommended by the distributors of analytical instruments like portable XRF is just not adequate when it comes to application on natural history specimens. But using numerous sample spots across a specimens body can really help in accounting for variation and better understanding of the mixtures inside each specimen. So in studying natural history specimens, the first task I was faced with is to narrow down the options because specimens come in a whole lot of different forms. So I chose to focus on mammals predominantly because a researcher called Borllati has recently shown that the distribution of endogenous pesticides and pollutants in plumage actually varies with the age of the feather rather than with the animal. And this in turn increases chemical variation across the specimen and makes differentiating between exogenous and endogenous accumulation of contaminants much more difficult to model. And I also chose study skins, which while they're more likely to have contamination from the storage environment and museum, they're also slightly less fiddled with chemically by the whole rehydration and mounting process. So I believe there may be less complex. I found I was able to create a statistically acceptable rate of variation, with only 18 spots on mammal skins compared to 45 spots on mounted bird skins. But my next round of data collection will reveal if I'm correct in all of that.

So guided by the 18th and 19th century instruction manuals, I made 90 replica skins using pig skin from the butcher. First I cleaned the flesh and the fat, I placed the skin in a bath. And then I began creating preservative mixtures and applying them in a manner that was both safe, but closely replicates application by handle brush. The skin was finally artificially aged to allow the preservatives to take effect. And these replicas have now formed a really good set of reference data, which I've been using to help disentangle the spectroscopic peaks generated by the chemical mixtures, and increased data accuracy overall. Then I applied a whole suite of non destructive spectroscopic techniques, to the skin, the pigskin specimens and also to seven specimens offered by the McClay collection at University of Sydney to see just what was achievable in analysis and what was realistically applicable in a museum context. Each stage, I refined the data collection techniques, I could reduce any form of damage on the specimens themselves. So I found two promising techniques. The first one is portable X ray Fluorescence (XRF) spectroscopy, and the second is Fourier Transform Near Infrared (FT-NIR) spectroscopy. And with a bit of additional support, both techniques allow the specimen to rest under their own weight while being analysed. Rather than being physically sampled or compressed under an instrument. Low Intensity X-rays pose little risk of damage to DNA, and low intensity and low wavelength infrared light has been shown in medical research to have very minimal impact on biological materials.

So bear with me, I'm gonna give you a brief introduction into the spectroscopic techniques. Portable XRF instruments emit an X-ray of fixed energy. And when that energy comes in contact with the specimen, it interacts with the atoms within it, and the X ray energy has changed and reflected back towards the instrument. The amount of this change is unique and characteristic for each element. The instrument collects all those energies and provides an output which indicates by the intensity of the peaks, the concentration of the elements within this specimen. So in one manoeuvre, I can find out about elemental composition, elemental distribution, and get an indication of the concentration inside the specimen. The Fourier Transform Near Infrared technique emits infrared light, and that makes the bonds within molecules begin to vibrate. And in doing so, the molecular bonds absorb some of that light energy. The instrument detects this change and outputs a graph that shows a unique fingerprint that can be used for identification and characterisation of the chemical structure within the specimen. The techniques work really well together. Because XRF is really helpful for inorganic components. And the infrared technique is really useful for organic components.

So while my research is still ongoing, with more than a year to go, I thought I'd show you a few interesting things that I've learned so far. At the beginning, what I was really worried about was that every specimen would look the same after 100 years or more in storage. But I was relieved to find that different preservatives, and different preservative recipes looked different in XRF. In fact, you can actually identify one museum specimen from another using the XRF spectrum. This comparison between five specimens from the Macleay collection shows differences both in the elements and in their concentration, massive relief. But it's all very well to know about the elements, but you can't really understand about preservatives without knowing about the organic components and organic components change over time. But using infrared techniques, you can see some of those components even now, and individual preservatives can be identified even after 100 years. So for example, in the Common Snipe, I was able to find a potassium sulfate and in a Rufus Bettong, calcium sulfate. And by combining a whole bunch of techniques together, you can get a really good profile of some of the preservatives that were used in a recipe. So in this Rufus Bettong we found iron sulfate, aluminium sulfate and a carbonised material that may have been burned prior to use. Some of the old recipes include a couple of preservatives that should be burned before application. And I'm wondering whether or not this carbonised material might be one of those.

Probably the most exciting thing is that it seems field collectors consistently used the same or very similar recipes. So when comparing 13 pademelons from the Australian Museum collection that were collected over four different expeditions, I found that Robert Grant specimens here in red, consistently look the same, and that Kendall Broadbent specimens in purple, also were consistent, and that they were both consistently different to each other. Broadbent used mercury and lead, copper and zinc in his preservative recipe. While Grant's recipe relied heavily on the preservative qualities of arsenic. This may indicate that field collectors can be differentiated using elemental analysis. But there's something else going on as well, because it appears that field collectors varied their methods. Maybe it's due to species or size or availability of materials. Here, three pademelons are compared with a possum or prepared by Grant. The dotted blue line shows that the possum is high chlorine, zinc and arsenic. I came across a really interesting note while sifting through documentation, the Macleay museum about two weeks ago, which suggests that the higher concentration of these elements may be an indicator for one particular bath preservative bath recipe. So that's the first rabbit hole I'll be looking into next year. And one of the other things that I found which is quite cool is that you can use all these multiple spots in XRF, to try and sort out the preservatives in pesticide application order. So by following co-occurring elements, I can see preservatives that were applied at the same time. So here in red, arsenic and mercury, align all the way along the animal which was suggested they were applied at the same time on this Kākāpō specimen.

So the next big thing for me is doing some principal component analysis and creating a model that allows the infrared and the XRF data to be integrated and offer a kind of more objective analysis of the data as a whole. I'm really keen to start applying the infrared techniques to those pademelons at the Australian Museum and start get some more detail on the preservatives used by Grant and Broadbent. Then data mining, to really draw some conclusions on preservatives, on specimens and on field collectors. And then last but not least, conduct a blind test to see if it's possible for me to pick a known field collector from a group of mystery specimens. And if it all works, I hope to share it all with others. So to finish up, I would like to thank my supervisors Dr. Carter, Dr. Lay and Dr. Philip and all the support team at Sydney University who have been great. Dr. Ingleby at the Australian Museum and the rest of the reconstructing museum data research team. Please feel free to give me your feedback and tell me what you thought was interesting and confusing. And feel free to ask any questions at all. Thank you so much for listening.