A preliminary assessment has been made of the application of amino acid racemization in molluscan fossils for chronostratigraphic studies of Late Quaternary coastal and marine sediments in South Australia. Evaluation of sources of amino acid D/L ratio variation provides a basis for interpreting the reliability of the technique. Several potential error sources have been recognized within the ambit of analytical and sampling error. In general, analytical error is low. In contrast, sampling error may give rise to a significantly larger margin of error, resulting from numerous factors unique to a deposit, or differences between deposits. The aminostratigraphic approach presents a convenient method of correlating strata and in establishing relative chronostratigraphies. This is illustrated in the correlation of the Late Pleistocene Glanville Formation and relative age assessments of Holocene and Pleistocene marine deposits from South Australia. Systematic variation of amino acid D/L ratios over latitudinally widely separated deposits of the Glanville Formation illustrates the temperature dependence of molluscan racemization and suggests similar latitudinal temperature gradients have existed from 120 ka BP to present. The species-specific nature of racemization is demonstrated, based on fauna obtained from the Glanville Formation. Results from age-calibrated Pleistocene marine deposits in South Australia suggest that amino acid racemization will serve for correlation of much of the Quaternary for temperate regions in Australia. �� Taylor & Francis Ltd.