Studies of cultured bovine aortic endothelial cells using quantitative chemiluminescence techniques have shown that the amount of nitric oxide released under basal conditions, or in response to either bradykinin or the calcium ionophore A23187 is insufficient to account for the vasorelaxant activities of the endothelium-derived relaxing factor (EDRF) derived from the same source. This observation contradicts previous suggestions that nitric oxide and EDRF are the same compound, but may be explained if EDRF is a compound that contains nitric oxide within its structure but is a much more potent vasodilator than nitric oxide. Such a molecule could be one of several nitrosothiols which may yield nitric oxide after a one-electron reduction. The present experiments were carried out to test the possibility that the biological activities of the endothelium-derived relaxing factor might more closely resemble those of one of these compounds, S-nitrosocysteine, than nitric oxide. Nitric oxide release from cultured bovine aortic endothelial cells was detected by chemiluminescence and bioassay experiments compared the vasodilator potencies of nitric oxide, S-nitrosocysteine, and EDRF. The results suggest that EDRF is much more likely to be a nitrosylated compound such as a nitrosothiol than authentic nitric oxide.