1 We examined whether beta(3)- and/or atypical beta-adrenoceptors relax the rat isolated mesenteric artery. 2 Mesenteric arteries precontracted with phenylephrine were relaxed by beta-agonists with the following potencies (pD(2)): nonselective agonist isoprenaline (6.00) > nonconventional partial agonist cyanopindolol (5.45) > beta(2)-agonist fenoterol (4.98) > nonconventional partial agonist CGP 12177 (4.19) > beta(3) agonist ZD 2079 (3.72). The beta(3)-agonist CL 316243 1 mm relaxed the vessel only marginally. 3 The concentration-response curves (CRCs) for cyanopindolol, CGP 12177 and ZD 2079 were not affected by the nonselective beta-antagonist propranolol 0.3 muM, the beta(2)-antagonist ICI 118551 1 muM and by CL 316243 60 muM, but shifted to the right by bupranolol (pA(2) 5.3-5.7), CGP 20712 (5.4) and SR 59230A (6.5-6.7) (the latter three drugs block atypical and/or beta(3)-adrenoceptors at high concentrations). 4 The CRC for isoprenaline was shifted to the right by propranolol (pA(2) 7.0) but, in the presence of propranolol 0.3 muM, not affected by SR 59230A 1 muM. The CRC for fenoterol was shifted to the right by propranolol (pA(2) 6.9) and ICI 118551 (6.8). 5 Removal of endothelium diminished the vasorelaxant effects of cyanopindolol, CGP 12177 and ZD 2079. 6 Fenoterol and cyanopindolol also relaxed (endothelium-intact) mesenteric arteries precontracted with serotonin. The relaxant effect of cyanopindolol was antagonized by bupranolol to about the same degree as in phenylephrine-contracted vessels. 7 In conclusion, beta(2)- and atypical beta-adrenoceptors (but not beta(3)-adrenoceptors) relax the rat mesenteric artery. The atypical beta-adrenoceptor, which is partially located endothelially, may differ from the low-affinity state of the beta(1)-adrenoceptor.