1. The vasoconstrictor activities of endothelin-2, endothelin-3, sarafotoxin S6b, human proendothelin1-38 and mouse vasoactive intestinal contractor (VIC) were studied in the isolated Krebs-Henseleit perfused mesenteric arterial bed of the rat in the presence and absence of the endothelium. The vasoconstrictor properties of endothelin-1 were studied in control preparations and in preparations treated with methylene blue or N omega-nitro-L-arginine methyl ester (NAME). Finally, the direct vasodilator properties of endothelin-2, endothelin-3 and sarafotoxin S6b were studied in preparations preconstricted with methoxamine. 2. In the presence of an intact endothelium, all of the peptides caused dose-dependent increases in perfusion pressure and sarafotoxin S6b was a full agonist relative to the other peptides studied (maximum increase in perfusion pressure, Rmax = 106 +/- 11 mmHg). Endothelin-1, endothelin-2 and VIC were more potent vasoconstrictors (ED50 93.0 +/- 40.0, 90.8 +/- 20.5 and 106 +/- 63 pmol, respectively) than endothelin-3 and sarafotoxin S6b, which were found to be equipotent (ED50 values 411 +/- 195 and 345 +/- 86 pmol, respectively). A full dose-response relationship could not be constructed for proendothelin, but the highest dose used (4 nmol) increased the perfusion pressure by 15.4 +/- 1.6 mmHg. 3. Destruction of the endothelium with the zwitterionic detergent 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulphonate (CHAPS) significantly enhanced the pressor activity of all 5 peptides. The Rmax for sarafotoxin S6b was not significantly altered by removal of the endothelium but its potency was significantly increased (ED50 = 115 +/- 15 pmol). Although their R,,, values were significantly increased, endothelin-2 and VIC were still partial agonists relative to sarafotoxin S6b in CHAPSpretreated preparations; their potencies were unchanged (ED5o values 118 + 53 and 416 + 196pmol, respectively). Removal of the endothelium significantly reduced the potency of endothelin-3 (ED5o, 6.3 + 2.2 nmol) but this peptide then exhibited full agonist activity (R..x = 106 + 14 mmHg). After endothelial cell destruction, the pressor responses to proendothelin were increased; 4 nmol gave a response of 38.8 + 5.5 mmHg. 4. Exposure of preparations to either 100 microM NAME (R,,,X = 42.6 + 2.4mmHg and EDSo = 57.5 + 13.7 pmol) or 10 microM methylene blue (R,,,. = 36.0 + 5.1 mmHg and ED50 = 81.5 + 26.1 pmol) significantly enhanced the maximum pressor responses to endothelin-l (control: R.,=X = 22.5 + 2.6 mmHg; ED5o = 93.0 + 40.Opmol). The values in the presence of NAME or methylene blue were not significantly different from those found previously for endothelin-1 after removal of the endothelium with CHAPS. 5. Endothelin-2, endothelin-3 and sarafotoxin S6b all caused vasorelaxation in preparations which had been precontracted with 100 microM methoxamine. This action was endothelium-dependent as it was abolished by perfusing the mesentery with CHAPS. Endothelin-3 and sarafotoxin S6b caused relaxation at much lower doses than were needed with endothelin-1 and endothelin-2. 6. The endothelium significantly modulates the vasoconstrictor activity of all the endothelin-like peptides studied, including the precursor peptide proendothelin (which was the least potent of the peptides). This modulation is likely to be due to the release of endothelium-derived relaxing factor, since similar results to destruction of the endothelium were obtained when endothelin-l was investigated in the presence of either methylene blue or NAME (an inhibitor of nitric oxide formation) in the perfusion fluid. The vasodilator effects of the peptides were also endothelium-dependent. There was a different order of potency for vasoconstriction and vasodilatation supporting the suggestion that there are sub-types of receptor for the endothelin-like peptides in the vasculature; one type on the vascular smooth muscle and a second type on the endothelium.