Faecal steroid hormone metabolites are becoming increasingly popular as parameters for reproductive functions and stress. The extraction of the steroids from the faecal matrix represents the initial step before quantification can be performed. The steroid metabolites present in the faecal matrix are of varying polarity and composition, so selection of a proper extraction procedure is essential. There have been some studies to address this complex but often neglected point. Radiolabeled steroids (e.g. cortisol or progesterone) have frequently been added to faecal samples to estimate the efficiency of the extraction procedures used. However, native, unmetabolized steroids are normally not present in the faeces and therefore the results are artificial and do not accurately reflect the actual recoveries of the substances of interest. In this respect, recovery experiments based on faecal samples from radiometabolism studies are more informative. In these samples, the metabolite content accurately reflects the mixture of metabolites present in the given species. As a result, it is possible to evaluate different extraction methods for use with faecal samples. We present studies on sheep, horses, pigs, hares and dogs that utilized samples containing naturally metabolized, 14C-labelled steroids. We recommend extracting faecal steroids by simply suspending the faeces in a high percentage of a primary alcohol (for glucocorticoid metabolites 80% aqueous methanol proved best suited for virtually all mammalian species tested so far). Not only does the procedure significantly increase the total amount of recovered radioactivity, it also increases the percentage of unconjugated metabolites, which are more likely to be recognized by the antibodies used in various immunoassays. The advantages of this extraction procedure are clear: it is very easy to use (no evaporation step is needed), it yields high recoveries and variation based on the extraction procedure is reduced to a minimum.