Abstract

The stratigraphic integrity of archaeological deposits in the seasonally wet tropics can be disrupted by termite activity and other processes. Significant questions have been raised about the surprisingly old age estimates for artefacts recovered from two sites in northern Australia: Nauwalabila and Madjedbebe. If accurate, a 65,000-year date for the latter would represent a 30% increase in the currently accepted length of human occupation of the continent. The oldest estimate for Nauwalabila falls just short of that. These dates would have important implications for human dispersal across Eurasia in the Late Pleistocene. No other data from Australia support them. Smith et al. (2020) claim to refute the notion that termite disturbance accounts for these anomalies. Here we show that the criteria on which the assertion is based are invalid. Radiocarbon data from both sites, not cited by Smith et al. (2020), are consistent with the termite-disturbance hypothesis. Luminescence data claimed to offer strong support for the early Madjedbebe estimate are also consistent with termite disruption. We conclude that the early dates for human presence at Madjedbebe and Nauwalabila must be rejected along with any chronometric inference about human behaviour to the degree it relies on them.