Stimulation of the endothelial lining of arteries with acetylcholine results in the release of a diffusible substance that relaxes and hyperpolarizes the underlying smooth muscle. Nitric oxide (NO) has been a candidate for this substance, termed endothelium-derived relaxing factor. But there are several observations that argue against the involvement of NO in acetylcholine-induced hyperpolarization. First, exogenous NO has no effect on the membrane potential of canine mesenteric arteries. Second, although haemoglobin (believed to bind and inactivate NO (refs 11-15)) and methylene blue (which prevents the stimulation of guanylate cyclase) inhibit relaxation, neither has an effect on hyperpolarization. Finally, nitroprusside, thought to generate NO in vascular smooth muscle, relaxes rat aorta without increasing rubidium efflux. Nevertheless, nitrovasodilators, nitroprusside and nitroglycerin cause hyperpolarization in some arteries. NO might therefore be responsible for at least part of the hyperpolarization induced by acetylcholine. We now report that hyperpolarization and relaxation evoked by acetylcholine are reduced by NG-monomethyl-L-arginine, an inhibitor of NO biosynthesis from L-arginine. Thus NO derived from the endothelium can cause hyperpolarization of vascular smooth muscle, which might also contribute to relaxation by closing voltage-dependent calcium channels. Our findings raise the possibility that hyperpolarization might be a component of NO signal transduction in neurons or inflammatory cells.