The secretion of vasopressin has been shown recently to be influenced by gonadal steroids. To further evaluate the relevance of sex steroids in the control of vasopressin secretion, the vasopressin response to dehydration has been studied in normal, castrated and sex-steroid treated castrated male rats. Short-term (3 weeks) castration did not modify the vasopressin response to 48 h of water deprivation. Long-term (10 weeks) castration, on the other hand, consistently reduced the vasopressin response by more than 50%. In both cases, the osmolality was unaffected. Treatment of long-term castrated rats with testosterone, dihydrotestosterone, or oestradiol increased the vasopressin response to dehydration towards normal control levels. However, only oestradiol could restore it to and beyond normal levels, although all three steroids were given in doses that were equipotent in bringing back to normal the plasma LH levels of the castrated animals. Angiotensin-II generation may indirectly be augmented by oestradiol treatment and this may account for the effect of oestradiol here reported. No such mechanism, however, may be invoked in the case of androgens; a direct modulatory effect at the hypothalamic level is postulated for explaining their influence on vasopressin secretion.