Acute glucoprivation profoundly stimulated hypothalamic-pituitary-adrenocortical (HPA) and adrenomedullary outflows. Whether these responses reflect a single central mechanism regulated by corticotropin-releasing hormone (CRH) has been unclear. This study examined the role of endogenous CRH in HPA and adrenomedullary responses to hypoglycemia in Sprague-Dawley rats, by using anti-CRH immune serum or a CRH antagonist (alpha-helical h/r CRH9-41, and in Lewis rats, a strain characterized by deficient hypothalamic CRH responses during stress. In conscious Sprague-Dawley rats with indwelling arterial and venous cannulas, insulin (0.3 U/kg was injected iv, and responses of serum glucose concentrations and plasma levels of corticotropin (ACTH) and catechols (including epinephrine, EPI; norepinephrine, NE; dihydroxyphenylalanine, DOPA; dihydroxyphenylglycol, DHPG; and dihydroxyphenylacetic acid, DOPAC) were assessed, with or without pretreatment with anti-CRH immune serum (0.5 or 1.0 ml iv or 10 microl icv) or alpha-helical h/r CRH9-41 (130 nmol iv or 13 nmol icv). Responses to insulin (1.0 U/kg iv) were also measured in conscious juvenile Lewis and Fischer 344/N rats. Insulin-induced hypoglycemia markedly increased plasma levels of EPI and ACTH in all groups. Pretreatment iv with 1/0 ml of anti-CRH immune serum blocked the ACTH response to insulin but failed to attenuate the EPI response. alpha-helical h/r CRH9-41, whether given iv or icv, failed to alter ACTH or EPI responses to insulin, although the antagonist did block EPI responses to icv CRH. Hypoglycemia elicited similar increments in ACTH levels in Lewis rats and Fischer 344/N control rats; and although Lewis rats had lower baseline EPI and smaller responses of NE, DHPG, DOPA, and DOPAC levels, the groups did not differ in proportionate increments in EPI levels. The results indicate that the ACTH response to hypoglycemia depends on availability of CRH outside the blood-brain barrier--presumably in the pituitary gland. The findings with icv alpha-helical h/r CRH9-41 can be explained by failure of the antagonist to reach effective concentrations at central sites of action of endogenous CRH, or by mechanisms other than CRH release determining the adrenomedullary response to hypoglycemia. Lewis rats seem to have less adrenomedullary secretion at baseline and smaller responses of NE synthesis and release during hypoglycemia than do Fischer 344/N rats. Neurochemical evidence for differential adrenomedullary and sympathoneural responses during hypoglycemia in all three rat strains is inconsistent with Cannon's view of a functionally unitary sympathoadrenal system. Lewis rats have deficient CRH responses to some stressors but not to others, or else pituitary-adrenomedullary responses in this setting depend on mechanisms other than CRH release in the brain. Both explanations are inconsistent with the doctrine of non-specificity, the main tenet of Selye's stress theory.