Electron spin resonance dating of South Australian megafauna sites

In recent years, there has been much debate about the timing of the Late Pleistocene extinction of the Australian megafauna. Some studies postulated a rapid, continental-wide extinction at around 46 000 years ago and that the arrival of humans in Australia, rather than climatic fluctuations, was the main cause for the demise of the megafauna. We have applied electron spin resonance (ESR) dating to a series of teeth from megafauna sites in South Australia, where young ages were expected. During this study, a number of unexpected problems were encountered. These were mainly related to the fact that ESR age assessments on fossil teeth are critically dependent on a realistic reconstruction of the post-mortem uranium uptake into the dental tissues. At virtually all sites, conventional, routine ESR dating, based on the parametric early and linear U-uptake models, would have led to grossly erroneous results. Most teeth were analysed for U-series isotopes with laser ablation ICP-MS, and the results were used to calculate combined U-series ESR age estimates. Only one of the 24 teeth analysed conformed to the commonly applied early and linear U-uptake models. At one of the sites, we found for the first time clear evidence of uranium leaching from dentine. Detailed laser ablation scans revealed that, in contrast to large mammals outside Australia, marsupial tooth enamel does not seem to contain a barrier layer close to the outer surface that blocks uranium diffusion into the enamel. As a consequence, uranium migrates into the enamel layer from both the outside and the inside via the dentine, which makes marsupial teeth generally less well suited for ESR dating. It was particularly difficult to obtain age estimates for the site of Black Creek Swamp, where the sediments contained extreme U-series disequilibrium, and the teeth had unexpectedly accumulated very high U-concentrations (up to 700 ppm in dentine) within a few thousand years. Although most of the sites contained reworked teeth, none of the samples yielded age estimates that were significantly younger than the proposed extinction window of about 40 000-51 000 years