1. We have used the isolated buffer-perfused superior mesenteric arterial bed of the rat to assess the involvement of ATP-sensitive potassium (KATP) channels in the vasorelaxant responses to beta-adrenoceptor agonists. 2. The vasorelaxant potencies of the non-selective beta-adrenoceptor agonist, isoprenaline, the beta 1-adrenoceptor agonist, dobutamine and the beta 2-adrenoceptor agonist, terbutaline were all significantly (P < 0.05) reduced (isoprenaline, ED50 = 265 +/- 31 pmol v. 1.05 +/- 0.42 nmol; dobutamine, ED50 = 294 +/- 67 pmol v. 497 +/- 115 pmol; terbutaline, ED50 = 157 +/- 26 nmol v. 452 +/- 120 nmol) in the presence of the KATP-channel blocker, glibenclamide. 3. The presence of glibenclamide only weakly influenced the vasorelaxant properties of salbutamol, a beta 2-adrenoceptor agonist, while those of verapamil, a beta-adrenoceptor-independent vasorelaxant, were unaffected. 4. In radioligand binding experiments, glibenclamide (1 nM-100 microM) did not displace any specific [3H]-dihydroalprenolol binding from rat beta-adrenoceptors. Therefore, glibenclamide does not bind to beta-adrenoceptors at the concentration used in the present investigation. 5. Vasorelaxant responses to dibutyryl cyclic AMP, the cell permeable analogue of cyclic AMP, were also unaffected by glibenclamide, indicating that the coupling of beta-adrenoceptors to KATP-channels occurs independently of the elevation of intracellular cyclic AMP. 6. We have shown that a significant element of the vasorelaxant responses to both beta 1- and beta 2-adrenoceptor activation involves the opening of KATP-channels. In conclusion, KATP-channels may play a physiological role in beta-adrenoceptor-mediated vasodilatation.