A large number and variety of compounds (acetylcholine, adenosine diphosphate, adenosine triphosphate, arachidonic acid, bradykinin, Ca2+ ionophores, calcitonin gene-related peptide, histamine, hydralazine, substance P, thrombin, and vasoactive intestinal polypeptide) have been shown to relax arterial smooth muscle indirectly. The endothelium in muscular arteries from several species appears to have receptors for these vasodilators. Binding of one of these compounds to its endothelial receptors results in the release (and presumably synthesis) of substance(s) that act on arterial smooth muscle to cause relaxation. The name endothelium-derived relaxing factor (EDRF) has been proposed for the substance or substances responsible for inhibition of contraction. Studies to determine additivity of endothelium-dependent relaxing agents and sensitivity of EDRF-mediated responses to a variety of inhibitors suggest that a single factor or a single common mechanism induces relaxation of vascular smooth muscle. Pharmacological studies have been equivocal with regard to the postulated involvement of phospholipases or arachidonic acid and to the suggestion that EDRF is an oxidative, non-cyclooxygenase product of arachidonate. Experiments on transfer of EDRF and reversal of endothelium-dependent relaxation consistently indicate that EDRF is quite labile. There is convincing evidence that EDRF activates smooth muscle guanylate cyclase, which results in an increase in intracellular cyclic guanosine 3',5'-monophosphate levels. The stimulation of guanylate cyclase by EDRF provides a valuable and sensitive parameter for studies with arteries as well as cells in culture. At present, the identity of EDRF and its role in cardiovascular homeostasis are unknown.