Abstract In this study, we have determined the distribution of corticotropin-releasing factor and vasopressin in the human hypothalamus, and investigated the effect of glucocorticoid administration on the concentrations of both peptides. Corticotropin-releasing factor and vasopressin were measured by a two-site immunoradiometric assay and/or radioimmunoassay. The presence of both peptides was studied in extracts of eleven areas of the human hypothalamus as well as in the pituitary stalk from autopsied patients who had been free of chronic steroid administration (n = 14) or had received Corticosteroids (n = 5). Unlike vasopressin, corticotropin-releasing factor was detected in all extracts: the highest concentration was found in the pituitary stalk, whilst the lowest detectable amounts occurred in the supraoptic and lateral areas and in the mammillary bodies. This pattern of distribution is similar to that reported for the rat hypothalamus. The excellent correlation (R = 0.994) between corticotropin-releasing factor data obtained by immunoradiometric assay and by radioimmunoassay renders the presence of a corticotropin-releasing factor precursor molecule in the extracts highly unlikely. In the human brain extracts, glucocorticoid treatment affected neither the content, nor the distribution of corticotropin-releasing factor and vasopressin. In the rats, dexamethasone administration produced a 50% decrease in the vasopressin content (P < 0.05) of the basomedial and dorsal parts of the hypothalamus and had no effect on the corticotropin-releasing factor content of these areas. These results show that the distribution of corticotropin-releasing factor is similar in both human and rat hypothalami. The rat data suggest that negative feedback effects of glucocorticoids involve changes in hypothalamic vasopressin content.