1. The isolated superior mesenteric arterial bed of the rat, perfused with Krebs-Henseleit solution containing 10 microM indomethacin, was used to study the effects of reducing dissolved O2 tension on the pressor responses to endothelin-1, endothelin-3 and sarafotoxin S6b. The modulation of these responses by the endothelium was investigated by removing the intima with the detergent CHAPS and, for endothelin-1, by inhibiting nitric oxide production with N omega-nitro-L-arginine methyl ester (L-NAME). Comparison was made with the effects of lowering O2 tension on the pressor responses to noradrenaline and 5-hydroxytryptamine. 2. Lowering the perfusate O2 tension from 551 +/- 2 mmHg to 14.0 +/- 0.5 mmHg did not change the ED50 for endothelin-1 but its maximal responses (Rmax) were increased by 2.1 and 2.7 fold, respectively, in the presence and absence of endothelium. The Rmax values for endothelin-3 were also greater in hypoxia either in the presence (by 2.3 fold) or absence of the endothelium (by 1.6 times) but those for sarafotoxin S6b were only enhanced significantly by hypoxia in the absence of the intima. hypoxia reduced the potencies of endothelin-3 and sarafotoxin S6b whether or not endothelium was present. 3. Endothelial destruction, whether in hypoxic or oxygenated conditions, increased the Rmax values for endothelin-1 and endothelin-3; at both O2 tensions those for endothelin-3 increased more than those for endothelin-1. The ED50 for endothelin-1 was unchanged by destroying the endothelium but endothelin-3 was less potent in the absence of an endothelium than in its presence. Removal of the endothelium did not change the R.ax for sarafotoxin S6b but increased its potency in both hypoxic and oxygenated tissues. 4. In hypoxia, and in the presence of both the endothelium and 100 microM L-NAME, the Rmax for endothelin-1 was 1.6 times greater than that in hypoxia in the absence of L-NAME. Co-infusion of 100 microM L-arginine, but not of 100 mircoM D-arginine, with 100 microM L-NAME reversed this effect. The presence of L-NAME decreased the potency of endothelin-1. 5. Destroying the endothelium did not affect the Rmax for noradrenaline in either oxygenated conditions or hypoxia. Changing 02 tension when the endothelium was intact had no effect on the Rmax but it was 11% greater in oxygenated, than in hypoxic, endothelium denuded preparations. Endothelial destruction decreased the potency of noradrenaline in hypoxia but increased it in oxygenated tissues. In hypoxia, L-NAME had no effect on the ED50 relative to control preparations with endothelium but the Rmax was 30% greater. 6. 5-Hydroxytryptamine gave very small pressor responses in the presence of endothelium in both oxygenated and hypoxic tissues but the Rmax was 1.7 times greater in hypoxia. L-NAME increased the R,,x by 9.8 times in oxygenated preparations and 6.3 fold in hypoxia. The ED5o values were the same in all conditions. 7. It is concluded that, although hypoxia generally increased the R.. for the endothelin/sarafotoxin peptides, the changes could not be explained by a simple increase in receptor number since hypoxia decreased the potency of endothelin-3 and sarafotoxin S6b. Thus alterations in receptor binding or activation properties, or both, also occurred. The changes associated with hypoxia were not common to all vasoconstrictor agonists since, in the absence of endothelial function, hypoxia did not affect the Rmax values for either noradrenaline or 5-hydroxytryptamine. Also, the pressor responses to the peptides and both the amines can be modulated by the endothelium in hypoxia as well as in oxygenated conditions.