Speleothems are now regarded as valuable archives of climatic conditions on the continents, offering a number of advantages relative to other continental climate proxy recorders such as lake sediments and peat cores. They are ideal materials for precise U-series dating, yielding ages in calendar years, thereby circumventing the radiocarbon calibration problems associated with most other continental records. Stable isotope studies in speleothems have shifted away from attempting to provide palaeo-temperature reconstructions to the attainable goal of providing precise estimates for the timing and duration of major O isotope-defined climatic events characterised by high signal to noise ratios (e.g. glacial/interglacial transitions, Dansgaard-Oeschger oscillations, the '8200-year' event). Unlike the marine records, speleothem data sets are not 'tuned', and their independent chronology offers opportunities to critically assess leads and lags in the climate system, that in turn can provide important insights into forcing and feedback mechanisms. Improved procedures for the extraction and measurement of stable isotope ratios in fluid inclusions trapped in speleothems are likely to provide, in the near future, a much enhanced basis for the quantitative interpretation of O isotope ratios in speleothem calcite. The latter developments open up once again the tantalising prospect of palaeo-temperature estimates, but more importantly perhaps, provide a direct test for a new generation of general circulation models whose hydrological cycles will incorporate the 'water isotopes'. The literature is reviewed briefly to provide for the reader a sense of the current state-of-the-art, and to provide some pointers for future research directions.