1. The aims of this study were to compare, in the rat isolated perfused lung preparation, the antagonist effects of a nonselective beta-adrenoceptor agonist (isoprenaline), a selective beta2-adrenoceptor agonist (salbutamol) and a selective beta3-adrenoceptor agonist (SR 59104A) on the hypoxic pulmonary pressure response, and to investigate the role of K+ channels, endothelium derived relaxing factor and prostaglandins in these effects. K+ channels were inhibited by glibenclamide, charybdotoxin or apamin, NO synthase and cyclo-oxygenase were inhibited by N(G)-nitro-L-arginine methyl ester (L-NAME) and indomethacin, respectively. 2. Hypoxic ventilation produced a significant increase in perfusion pressure (+65%, P<0.001) and L-NAME significantly increased this response further (+123%, P<0.01). After apamin, L-NAME, indomethacin, post-hypoxic basal pressure did not return to baseline values (P<0.001). 3. Glibenclamide partially inhibited the relaxant effects of isoprenaline (P<0.05) and salbutamol (P<0.001) but not that of SR 59104A. In contrast, charybdotoxin and apamin partially inhibited the relaxant effects of SR 59104A (P=0.053 and <0.01, respectively) but did not modify the effects of isoprenaline and salbutamol. L-NAME partially inhibited the dilator response of salbutamol (P<0.01) and SR 59104A (P<0.05) but not that of isoprenaline. 4. We conclude that (a) EDRF exerts a significant inhibition of the hypoxic pulmonary response, (b) SK(Ca) channel activation, EDRF and prostaglandins contribute to the reversal of the hypoxic pressure response, (c) the vasodilation induced by isoprenaline is mediated in part by activation of K(ATP) channels, that of salbutamol by activation of K(ATP) channels and EDRF. In contrast, SR 59104A partly operates through BK(Ca), SK(Ca), channels and EDRF activation, differing in this from the beta1 and beta2-adrenoceptor agonists.