Despite Diprotodon fossils occurring widely across Australia, until recently, few finds hake been adequately dated. This is due to several reasons, primarily the inadequacies of the radiocarbon methods. New dating methods, which coincidentally increase the datable age range, have been developed in recent years. One of these is thermoluminescence (TL) dating. Yet there are still few reliably dated Diprotodon specimens because they, must be found and dated in situ. A chance discovery in 1992 gave the authors an opportunity, to test one of these ne,A methods and at the same time solve a thirty year old mystery. An articulated portion of a Diprotodon skeleton found at Hallett Cove is associated kith sediment TL-dated to about 55 000 years, and is also a possible source for a fossil tooth found on the nearby beach in 1971.

Determination of geological/archaeological ages by luminescence dating requires the measurement of the rate of delivery of radiation dose from the sample and its environment. For all age determinations it is necessary to assess whether the dose-rate has been constant during the time that the dose has been accumulating, and to make allowance for it if it was not. We report here results from a number of Australian sites to illustrate the information that can be found about element concentrations, about their distribution and about dose rates. Tests for radioactive disequilibrium form part of these measurements. Such tests take the form of comparison of independent measurements of dose-rates and/or individual elements. In general, there is little or no evidence for disequilibrium in aeolian dune systems, even when the age of the system exceeds 500 ka. However, disequilibrium has been found in several sites subject to wet conditions.

In previous work, we have discussed the way in which the contribution of cosmic rays to the dose rates for TL/ESR dating depends on altitude, latitude and depth below ground level. The present paper extends the discussion to greater depths. Long-term changes in the geomagnetic field and the galactic cosmic ray intensity could affect the dose rate. The effects are shown to be small over the past 500 ka in most circumstances.

Our thermoluminescence dating technique has been applied to several new samples from the remarkable sequence of stranded beach dunes in southeast South Australia that were formed during the past 800 ka. We show that Robe III is probably 100 ka as suggested by Schwebel and not 80 ka as we suggested earlier. A sample from the West Dairy dune is shown to belong to δ18O Stage 7 in accordance with Schwebel's original suggestion and not Stage 5 as he suggested later. A new West Naracoorte sample yielded acceptable data and an age consistent with formation just after the Brunhes-Matuyama magnetic reversal at 780 ka as required by palaeomagnetic measurements. Two more samples from the Woakwine dune show that our reproducibility is consistent with our derived errors.

Alkali feldspars, either extracted or as a component of an undifferentiated mixture, are important in thermoluminescence and optical dating. The present work describes measurements of three-dimensional thermoluminescence spectra, in which emitted light intensity is displayed as a function of both temperature and photon energy (or wavelength), for feldspars of a range of composition. In consideration of the increased role of TL/OSL in dating sediments, a particular emphasis has been an assessment of the way in which bleaching by visible and ultra-violet light affects the spectrum. Generally speaking, prolonged bleaching is necessary to remove TL from feldspars but most high-sodium feldspars in the study have a readily bleached component in the ultra-violet emission. We have not found spectral evidence for any other readily bleached peak at high glow-curve temperatures similar to the 325°C peak found in quartz. For some samples, bleaching results in increases of emission in the temperature region of interest for TL dating, and these increases appear across the emission spectrum.

Thermoluminescence (TL) dating is a useful tool for determining the age of prehistoric earthquakes by dating deposits that are stratigraphically related to fault scarps that formed during the earthquakes. TL dating of aeolian sand in the area of the 1988 Tennant Creek, Northern Territory, earthquakes provides evidence that similar earthquakes have not ruptured the causative faults for at least 50 ka. Pilot TL measurements of deposits associated with the Roopena and Ash Ridge fault scarps near Whyalla on Eyre Peninsula, South Australia, suggest an age of 140 ka for the Quaternary deposits associated with the formation of the scarps.

Feldspars occur in many sediments, and dating them by thermoluminescence (TL) requires a knowledge of how efficiently sunlight bleaches these minerals. The relative bleaching efficiencies of various wavelengths have been investigated for four alkali feldspars and an oligoclase, using the natural TL of the samples and the artificial TL induced by laboratory irradiation, bleached by wavelengths in the range 322–350 nm. Ultraviolet wavelengths are more efficient than the longer wavelengths at bleaching all the glow-curve peaks and are responsible for most of the phototransfer from high-temperature peaks to low-temperature peaks. Some feldspars show an initial increase in TL intensity with bleaching. Two sanidines of similar bulk composition but different TL glow curves different dependences on bleaching wavelength. In a sodic sanidine with a dominant 240°C peak the energy required to bleach to 50% was a thousand times greater at 550 nm than at 322 nm. In contrast, another sanidine with a complex glow curve between 150 and 450°C showed a much slower change in the 330°C peak with wavelength—a factor of 10 between 322 and 550 nm—with a region between 400 and 500 nm which was almost independent of wavelength. This behaviour is explained by competing mechanisms of phototransfer and bleaching which depend on wavelength in different ways.

The remarkable sequence of stranded beach dunes in south-east South Australia deposited during the past 800 ka has provided us with a rare set of samples for testing thermoluminescence sediment-dating methods. Here we report thermoluminescence ages obtained from quartz separated from the dunes, and compare them with accepted ages for the high sea-levels responsible for the dunes. The independent ages of the dunes were obtained from earlier work by Schwebel which has been extended and refined by modelling the dune formation using sea-level variations derived from the oceanic deltaO-18 record. The work is unusual in extending thermoluminescence dating to an 800 ka sequence with modest geological control of the age. The agreement' between thermoluminescence and geological ages is satisfactory for eight dunes with ages in the range 120-800 ka.Quartz is not usually thought to be useful for thermoluminescence dating in this time span; two factors contribute to our success. The first is that the dose rates are low, about 0.5 Gy ka-1, or one quarter of that typical of sediments. The second is that the thermoluminescence vs. dose response continues to rise monotonically above the saturation region at doses above 300 Gy.

A sequence of stranded beach dunes in the southeast of South Australia which resulted from the advance and retreat of the sea over a tectonically rising land surface during the past 800 ka has been securely dated by geological means and successfully used to test thermoluminescence dating based on quartz. We have now obtained a series of promising results from the same sequence with a test of optical dating on inclusions within the quartz grains. The innovative aspect of the measurements is the use of infrared irradiation to stimulate emission from the inclusions rather than conventional stimulation of the quartz itself by light of shorter wavelength. Satisfactory ages were found for seven dunes covering the time span 0-400 ka.

Thermoluminescence, photoluminescence and electron-spin resonance dating require the accurate determination of the ionizing radiation dose rate experienced by the sample. The availability of portable multichannel analysers with calculation and storage facilities, gives an immediate read-out of total γ-ray dose rate, potassium, thorium and uranium concentrations and provides storage for spectra. Extensive measurements have been made for sites at which disequilibrium is not expected and the results agree very closely with those obtained by XRF and neutron activation. For potassium and thorium, the field measurement is accurate and precise and for uranium the measurement is also accurate if there is no disequilibrium. Any discrepancy with a non-radiometric determination of uranium suggests disequilibrium and the necessity for further investigation.

Feldspars are an important component of materials used for thermoluminescence (TL) and photoluminescence (PL) dating of sediments; and successful dating implies a knowledge of the degree of re-setting by exposure to sunlight of the stored luminescence energy. We have studied the bleaching by full sunlight of the TL of six alkali feldspars of representative composition and of one oligoclase. The high potassium and high sodium feldspars are the brightest, and are also the least easily bleached, whereas those of intermediate (K-Na) composition bleach quickly. On a time scale of upwards of 16 h, all samples would have bleached enough for them to be useful for TL dating. A search was made for individual glow curve peaks that bleach particularly quickly (and are thus similar to the well-known 325°C peak used in the TL dating of quartz). In the samples studied, there was only one glow curve peak that was selectively susceptible to bleaching: the 280°C peak in oligoclase.

Thermogravimetric and multi-element analyses of archaeological human rib and modem mammal controls from the lower Murray River region of South Australia indicate that magnesium, strontium, and barium concentrations in the archaeological bone have been altered significantly by post-mortem chemical reactions in the burial environment. Ionic exchanges between soil solution and buried bone resulted in lower bone magnesium concentrations and higher bone strontium and barium concentrations. Much of the remaining magnesium is associated with diagenetic calcite. Although the archaeological bone exhibits increased hydroxyapatite crystal size, the strontium and barium increases are not correlated with crystal growth. The complex interactions of these post-depositional chemical processes will complicate attempts to reconstruct in vivo elemental values from archaeological bone.

The formation of shoreline gypseous dunes is a major event in the Quaternary history of many playas in central Australia. The dunes probably were formed during a period of high regional water table when abundant gypsum was deposited in a near‐shore groundwater seepage zone and deflated on to the shoreline dunes. Ten samples from two sites at Lake Amadeus, a major playa in the region, provided coarse (90–125 μ) quartz grains for thermoluminescence (TL) dating. Both regeneration TL and additive TL were measured. A well defined regeneration curve is basically consistent with the additive curves. Three methods, alpha counting, analyses of radioactive isotope concentrations and on‐site gamma scintillometer measurements, were used to evaluate the dose rates, giving consistent results. Although the equivalent doses of the samples scatter from 25 Gy to 71 Gy, the TL ages of all samples from the two sites cluster in the range 45–60 ka.

Thermoluminescence (TL) of quartz grains has been used to date a soil horizon at each of four sites in a chronosequence of freely drained podzols at Cooloola and North Stradbroke Island. The chronological order of the TL dates is in agreement with the sequence of inferred ages based on stratigraphic, geomorphic, denudational and pedological evidence, but at least one of the TL dates is of considerably greater age than the field evidence implies. Possible explanations of this anomaly are discussed. Differences in the nature of the quartz grains and the various pre-treatments used in TL dating were also examined by using a combination of thermogravimetric analysis (TGA), differential thermal analysis (DTA) and scanning electron microscopy (SEM) methods. The DTA showed no dissimilarity among the quartz samples from the different sites, but the TGA results showed significant weight losses for some samples and pretreatments. The SEM method further showed that weight loss is associated with water-sorbing substances (allophanic materials) present as coatings on grain surfaces and/or in cracks within grains.

Accurate paleoecological interences from the elemental composition of human and faunal bone recovered from archaeological and paleontological sites will depend on controls for postmortem diagenesis. In vivo chemical signatures in bone can be obscured by postmortem changes in the soil environment. Models addressing the fossilization process in bone must consider both ionic substitution and mineralization mechanisms of diagenesis. Current solubility profile techniques for the differentiation of biogenic and diagenetic chemical phases in fossil bone neglect ionic substitution. Ionic soil solution data from an arid-land South Australian archaeological site are represented to demonstrate the importance of such information to the development and testing of these predictive models. Elemental rations in the soil solution are compared with those in rib bone from archaeological human skeletons and modern terrestrial mammal controls. The archaeological bone is enriched with Si, Al, Mn, Ba, Fe, S, Sr, K and Ti and depleted in Mg relative to the modern controls. A variety of processes including contamination with quartz and secondary carbonates and ionic substitution from the soil solution are responsible for these postmortem chemical alterations.

A new high-sensitivity spectrometer was used to study the thermoluminescence spectra of several quartzes. It was found that natural quartz extracted from sediments, but otherwise untreated or given a gamma dose, yields two broad emission bands, one centred at 630 nm and the other extending from 350 nm to at least 550 nm. The dependence of the spectra on radiation dose were examined for both laboratory dose and natural dose, which differ by 9 orders of magnitude in dose rate, and no effect due to dose rate was observed. The dose responses of both emission bands were examined in the range 0–1600 Gy and were found to be highly non-linear and different. It is shown that the predose effect in the high temperature thermoluminescence peak is associated with the blue emission and not the red emission.