Thirty years ago, a radiocarbon date of 'more than forty thousand years' from Lake Mungo became both a slogan for Indigenous pride and a challenging focus for archaeological science. It wasn't much of a date (and consequently never published) but everyone wanted to know for how much more than 40 000 years people had been living in Australia. New developments in numerical dating techniques are now providing answers to that question, and there is similar progress in the understanding of climate change, landscape evolution and extinction of the Late Pleistocene megafauna. This update is based on a review (Gillespie 2002) of the oldest archaeological sites from the Late Pleistocene continent of Australia, when Tasmania and New Guinea were connected to the mainland by land-bridges (Figure 1).
[FIGURE 1 OMITTED]
It should be remembered that dating laboratories do not actually measure the age of samples. Quite a lot of the work they do is analytical chemistry, measuring the concentrations of elements and isotopes that give rise to a time-varying signal in some component of a sample. Specialists in the numerical dating techniques use the analytical methods that they consider to be appropriate, and this primary data is converted into age-estimates via mathematical procedures. As the technology for both analytical chemistry and measurement of the signal of interest improves, providing higher precision on smaller samples, best practice for any given laboratory will also change--if they can afford the regular upgrades.
Developments in numerical dating methods
Radiocarbon analysis has been the dominant method for estimating the age of Holocene and Late Pleistocene archaeological and geological sites for more than 50 years. Three aspects of the radiocarbon method continue to tax practitioners and are of concern to archaeologists and others who rely on radiocarbon results: maximum possible age, contamination, and calibration. That famous result from John Mulvaney's Mungo excavation could not be distinguished from laboratory 'background' measurements. In the early 1970s, background for the ANU laboratory (and the similar University of Sydney laboratory) was about 40 000 BP, so nothing older than that limit could be determined. Improvements in sample processing and measuring equipment have now shifted the background for many laboratories to 45 000 BP and the best can manage 55 000 BP or more, so radiocarbon age-estimations greater than 40 000 BP certainly now are possible.
Most readers are familiar with 'contamination' of radiocarbon samples, usually by materials younger than the true sample age. Much improved chemical decontamination procedures have been made possible by the small-sample capability of accelerator mass spectrometry (AMS) systems. One example is the change from routine acid/base/acid (ABA) chemistry to acid/base/wet oxidation (ABOX) methods for the removal of humic acids from charcoal. Radiocarbon measurements on microfauna and flora, such as foraminifera and pollen, and specific molecules such as single amino acids are also possible. These procedures give more reliable results because the identity of samples being dated is more accurately known and contamination can be reduced to insignificant levels.
Radiocarbon analysis of annual tree rings shows deviations from the exactly known tree-ring age, so the idea of 'calibration' was developed whereby radiocarbon years are converted to calendar years. The data currently accepted by the radiocarbon community (known as INTCAL04) extends to 26 000 calendar years ago, but a consensus calibration curve back to 50 000 years should be available within two or three years (Bard et al. 2004). Calibration of radiocarbon age-estimates is necessary for comparison with other dating methods which yield dates in calendar years.
Thermoluminescence (TL) dating was developed for ceramics and still works well for that application because it relies on heating in antiquity to reset the TL signal to zero, but results for sediments that were never heated have been mixed. …