Alcohol administration activates the hypothalamic-pituitary-adrenal (HPA) axis of both male and female rats, with females secreting more adrenocorticotropin (ACTH) and corticosterone than males in response to the same dose of alcohol. Our earlier work suggested that this gender difference arises due to the activational effects of gonadal steroids. In particular, we hypothesized that both androgens and estrogens play a role, with androgens exerting an inhibitory influence while estrogens elevate activity of the HPA. In the present studies, we tested this hypothesis by manipulating steroidal milieu in male rats using surgical castration and chronic implantation of testosterone (T), dihydrotestosterone (DHT), or estradiol (E2). Intact male and female rats were included as controls. Injection of alcohol (3 g/kg b.wt., i.p.) resulted in elevation of blood alcohol levels, ACTH and corticosterone in all groups. However, the amount of ACTH secreted was greater in females and castrated males implanted with E2 than in intact males. In castrated males, regardless of androgen implantation, the ACTH response was intermediate, with mean levels between those of females and males, but not differing significantly from either. In contrast to the ACTH results, significantly higher corticosterone secretion was measured in females and castrated males which did not receive a steroid implant. Since there were no significant differences between groups in blood alcohol levels (BALs), these results are not due to steroid-dependent alterations in alcohol metabolism. Because the ACTH data confirmed an activational effect of E2, we sought to determine whether this steroid regulated levels of corticotropin-releasing factor (CRF) and arginine vasopressin (AVP) mRNAs in the paraventricular nucleus of the hypothalamus (PVN). Four pretreatment groups were studied: intact males, intact females, castrated males, and castrated males implanted with E2. Two weeks after surgery, alcohol or vehicle was administered 3 h before brains were collected. In intact males, alcohol treatment elevated levels of both CRF and AVP mRNAs in the PVN, as previously reported. In contrast, this treatment decreased CRF mRNA in castrated males implanted with E2. In addition, steroid pretreatment alone elevated CRF mRNA levels in castrated males. Although we did not observe E2-dependent increases in CRF or AVP mRNAs, our data do support a complex effect of gonadal steroids on expression of these mRNAs in the PVN.