The mining and quarrying industries can summon deep emotions when mentioned in polite conversation. Polar views might be offered by your friends and family, with words like “destructive” and “irreparable” countered with “employment” and “vital”. While mining issues carve deep ravines into the political landscape in my home town, some of the most productive fossil sites that I have visited were discovered by agricultural quarrying. Indeed, many fossil sites around the world have been discovered serendipitously by digger, tractor or shovel.
The mining industry has also contributed to analytical advances in fossil studies. X-ray fluorescence spectroscopy (XRF) is an analytical technique that uses x-rays to probe the chemical composition of a sample. This is traditionally performed in a dedicated lab, where samples are ground to flour, mixed with another powder, fused into glass, and analysed in a room-sized instrument. In stark contrast, a range of portable instruments are produced that can perform x-ray fluorescence spectroscopy in the field. These instruments have been optimised for the mining industry, providing a rapid way of targeting mineral-rich areas. The chemical mapping that is applied in the hunt for platinum and gold is conceptually identical to describing the heterogeneous burial environments of different fossils: elements, minerals and rocks vary from place to place. Following this logic, we used a handheld XRF instrument to describe the burial environments of fossil bones in the Western Cape of South Africa.
We studied fossil antelope bones and teeth from two Pleistocene localities, one that has always been located inland, and one that presently crops out along the coast (Thomas and Chinsamy, 2011). Bones from the inland site are iron rich, whereas bones from the coastal site feature an abundance of calcium (more than what you expect for just bone). Although this is a simple conclusion, we discovered that the chemical information provided by handheld XRF needs some careful interpretation. Several factors can influence handheld XRF data: the distance between the sample and the instrument, the density of the sample, the calibration of the in-built software. The most informative approach is to analyse the raw x-ray fluorescence spectra, not the elemental concentrations. Principal components analysis score values provide meaningful groups, and the corresponding loadings help to understand those groups. Our key finding from studying fossils with an instrument originally designed for mining exploration was that multivariate statistics are essential (unavoidable, indispensible, vital and crucial) for interpreting the dataset. Only once the chemometics are sorted will the chemical and paleontological story spring to life.
Thomas, D. B., Chinsamy, A. 2011. Chemometric analysis of EDXRF measurements from fossil bone. X-ray Spectrometry 40: 441–445