Age-related changes in glucocorticoid negative feedback inhibition of hypothalamic CRF and pituitary ACTH are observed in rodents. Attempts to study similar effects in humans have produced mixed results due in part to the difficulty in matching older subjects on social and lifestyle variables. The present study used female rhesus monkeys as a model for women by comparing young adult (n = 20) to old (n = 20) females to test the hypotheses that the hypothalamic-pituitary-adrenal axis is altered in older animals and that this difference is exacerbated by exposure to social stress. The effects of age on the response to two doses of dexamethasone and two doses of CRF were assessed in females living in a stable social environment (control) and in socially stressed females removed from their group and housed temporarily in a remote, nonsocial environment (separated). The suppression of serum cortisol was not different between the two doses of dexamethasone. Before dexamethasone administration (2100 h), serum cortisol was significantly higher in old control females than in either young or old separated females, who were not different from one another. The young control females had baseline cortisol concentrations significantly lower than all other groups. Serum cortisol was suppressed approximately 75% below baseline values in all groups by 10 h after dexamethasone administration. Age significantly affected serum cortisol after dexamethasone, as the old control group showed a release from suppression 19 h posttreatment compared to the young control group and compared to the separated groups. Social condition had a significant effect on the responses of serum cortisol and plasma ACTH to CRF administration. At baseline (0930 h), serum cortisol was significantly higher in young controls compared with older controls, with both separated groups having intermediate values. Similarly, plasma ACTH at baseline was significantly higher in young controls compared to all other groups. Social separation significantly diminished the elevation of both serum cortisol and ACTH after stimulation with either dose of CRF. Control females showed a prolonged increase in plasma ACTH through 60 min and an increase in serum cortisol through 120 min after CRF. In contrast, these hormones either declined by 60 min or did not increase in socially separated females after CRF administration. These data suggest that the circadian rhythm in serum cortisol may be affected by aging, as levels were higher in the evening and lower in the morning in old control compared to young control females. The effect of age on the response to dexamethasone treatment among the control groups lends support to the hypothesis that the sensitivity of glucocorticoid negative feedback diminishes with aging. Although age did not affect the response to CRF, social separation diminished the elevation in both serum cortisol and plasma ACTH. Whether this effect was due to stress-induced down-regulation of pituitary CRF receptors remains to be determined.