Benzodiazepines and the novel anxiolytic buspirone share a common capacity to relieve clinical anxiety but do not share any side effects. Anxiety releases stress hormones and, at moderate doses, anxiolytic benzodiazepines block this release. It is interesting, therefore, that buspirone and other 5-HT1A agonists release stress hormones at moderate doses. Both the U-shaped dose-response curve seen with buspirone in some animal tests of anxiety and its slow onset of clinical action could be attributed to this release of stress hormones. Metyrapone (200 mg/kg), an inhibitor of 11-beta-hydroxylase, was used in the present experiments as a form of chemical adrenalectomy and was combined with administration of corticosterone (1 mg) to produce rats with presumed approximately normal corticosterone levels but no capacity to release endogenous corticosterone. This treatment reduced the difference normally observed in the effects of chlordiazepoxide (5 mg/kg) and buspirone (0.37 mg/kg) on a fixed interval schedule particularly in the early part of the interval when release of behavioral inhibition would be expected to contribute most to the effects. These results are consistent with the previous suggestion of Johnston and File (8) that the anxiolytic action of buspirone may be counteracted by activation of the pituitary-adrenal axis. Corticosterone appears to be the most likely critical agent for this antagonist action in the present experiments, although CRF and ACTH are also possibilities. It is likely that there is a mutual functional opposition between endogenous anxiolytic factors and stress hormones.