Stable carbon isotope ratio analysis of carbonate substituted within fossil tooth enamel has been shown to be a useful approach to questions about diets and environments of extinct animals. Tooth enamel appears to be preferable to fossil bone apatite as sample material, at least when the material is of great age (>100 Kyr to Myr range). After appropriate pretreatment to remove contaminants, enamel carbon isotope ratios directly reflect dietary values and are relatively immune to chemical alteration. Fossil teeth, however, are very precious because they are readily identifiable and of great behavioural and taxonomic value. This has imposed severe constraints on wider application of isotopic techniques which have until recently required destructive analysis of relatively large amounts of enamel. Recent developments in mass spectrometry allow the measurement of very small gas samples, coupled with automated gas production and greater numbers of analyses. These developments carry obvious advantages and present new opportunities, but we encountered some unexpected problems in scaling down the techniques. We outline some changes required for scaling down pretreatment procedures for sample sizes about two orders of magnitude smaller than before. Tiny sample requirements are useful in several ways. Firstly we can address questions about intra-tooth variability, and inter-tooth variability along toothrows, as a way of estimating isotopic ranges for animals which may change their diets, perhaps seasonally or through time, and we present some preliminary data exploring these possibilities for a modern elephant and several baboons, respectively. Secondly, smaller sample requirements are particularly useful in the case of isotopic analysis of carnivores, both because of the relative rarity of their teeth and their thin enamel. We present some data which further explore carnivore diets and dietary shifts during the periods encompassed by Members 1 and 2 at Swartkrans Cave, South Africa.