1. Although it well established that corticotropin-releasing factor (CRF) injected i.v. can cause hypotension and vasodilatation, there is no in vitro evidence that CRF acts as a vasodilator. We have therefore tested the hypothesis that the hypotensive effect of i.v. CRF is due to a direct vasodilator action by carrying out experiments in vitro on rat resistance arteries (i.d. 150-300 microns). 2. Initial in vivo experiments confirmed that CRF (1.5 nmol.kg-1) injected i.v. caused hypotension in rats, this being partially antagonized by the CRF analogue CRF9-41. 3. For the in vitro experiments, vessels were taken from the mesenteric, cerebral and femoral vascular beds, and mounted as ring preparations in an isometric myograph. The vessels were pre-contracted with one of 3 agonists (prostaglandin F2 alpha, arginine vasopressin or noradrenaline) or with a high-potassium solution (K+). 4. With maximal concentrations of the agonists, CRF caused relaxation of mesenteric and cerebral vessels with 10 nM, and near complete relaxation with 100 nM. Femoral vessels pre-constricted with agonists and all vessels pre-constricted with K+ were less affected by CRF. In the mesenteric vessels, with sub-maximal levels of pre-constriction, CRF caused substantial relaxation at 1 nM and could cause complete relaxation at 10 nM. 5. The relaxant effect of CRF on contractions of mesenteric vessels was antagonized by 100 nM CRF9-41. Neither tetraethyl ammonium (30 mM) nor glibenclamide (3 microM) antagonized the relaxant effect of CRF. 6. The relaxant effect of CRF on mesenteric small arteries was found to be unaffected by removal of the endothelium. 7. The results indicate that CRF causes an endothelial-independent vasodilatation of rat resistance arteries under in vitro conditions at concentrations which are consistent with this being an important cause of the hypotension observed with i.v. injection of CRF.