Analyses of the natural abundances of stable isotopes of carbon, hydrogen, oxygen, nitrogen and sulphur have widespread applications for the natural sciences. In the short term, the relative depletion of '3C in plants has been used to model photosynthetic limitation and carboxylation pathway, and leads to the long-term analysis of canopy water-use efficiency, trophic levels in ecosystems and evaluation of changes in global CO2 supply. Alternatively, the abundance of D (2H) and 180 in waters, sediments and plants has been used to model climate change over millennia, although the initial fractionation processes in plants still require elucidation. While '5N and 34S are apparently not subject to such large fractionation during assimilation by primary producers, they can be used to deduce origins and sources of N and S fertilization in ecosystems. With a wealth of recent review material dealing with the specific applications of these isotopes (Kaplan, 1975; DeNiro, 1987; Peterson & Fry, 1987; Raven, 1987; Ehleringer & Osmond, 1989; Rundel, Ehleringer & Nagy, 1989), a brief review such as this must attempt to summarize and integrate recent developments. Given the exponential interest in stable isotope analyses, the aim of the following article should be: first, to highlight methodological and theoretical constraints for any new study; second, to evaluate the current applications in physiological ecology; and finally to suggest some potential developments. It is hoped that anyone planning an excursion into this field will find this approach useful, although apologies must be made for the subjectivity and brevity with which these applications have been selected and summarized and for the preponderance of the applications involving carbon. Where appropriate, reference will be made to more comprehensive studies which will enable the detailed background of each isotope to be investigated, although applications involving the addition of stable isotopes for 'labelling' are not considered.