SUMMARY We have shown previously that the capacity of the jejunal mucosa to oxidise testosterone to the weaker androgen, androstenedione, by the enzyme 17j3-hydroxysteroid dehydrogenase (1 7f3-HSD), is considerable. The present study extends these earlier observations by measuring 1 7,B-HSD activity in different regions of the gastrointestinal tract, by investigating the potential for testosterone metabolism by slices and everted sacs of rat jejunum, and estimating the contribution of intestinal testosterone metabolites to circulating levels of plasma androgens, by portal vein sampling in the rat, in vivo. 17f3-HSD activity in homogenates of gastric and duodenal mucosa was significantly higher than that in jejunum, and was also present in ileum and colon. In addition to androstenedione, slices and everted sacs of rat jejunum produced various metabolites, one of which was probably dihydrotestosterone. It was not, however, a major metabolite in vivo. It is suggested that 5a-reduction may be favoured in vitro by a lower oxidation-reduction potential resulting from tissue anoxia. The major portal vein metabolite was androstenedione, the same major metabolite produced by mucosal homogenates. We conclude that oxidation of testosterone is the major metabolic pathway in intestinal mucosa and the capacity of the gastrointestinal tract to reduce the potency of testosterone is considerable. Our findings suggest that the gut, rather than the liver, is responsible for the failure of oral testosterone to provide effective androgen replacement therapy. The qualitative difference in testosterone metabolism between in vitro and in vivo preparations emphasises the need for caution in the interpretation of similar in vitro experiments.In a previous study we demonstrated that homogenates of rat and human jejunal mucosa can oxidise testosterone to the weaker androgen androstenedione.' This conversion was independent of gut bacteria. The enzyme responsible for this conversion, 1 7f3-hydroxysteroid dehydrogenase (1 7f-HSD), was present at high specific activity and the overall capacity of the small intestine to perform this conversion appeared considerable. This suggested that the gut, like other nonendocrine tissues such as skeletal muscle,2' lung,4 and adipose tissue,' 6 may play a significant role in the peripheral metabolism of testosterone.Previous reports7"1' have indicated that mammalian intestinal mucosa is also capable of reducing testosterone in relatively minor amounts to a variety of 5a-reduced metabolites, including the more potent metab-