Reactive oxygen species (ROS) are important mediators of vascular tone and play a key role in cardiovascular disease. We showed previously that LY83583, a generator of superoxide anion, caused a superoxide dismutase (SOD)-inhibitable constriction of rat pulmonary arteries that was directed primarily via Rho-kinase-mediated Ca2+-sensitization (Knock et al., 2009). In the present study we investigated the vasomotor responses to LY83583 and hydrogen peroxide (H2O2) in intact and alpha-toxin permeabilised U46619 (100 nM)-pre-constricted rat mesenteric arteries (MA), mounted on a wire myograph. In intact MA, LY83583 was applied in the presence of L-NAME (1mM) in order to eliminate superoxide scavenging of NO. Statistical analysis was by Student t-test. Both LY83583 (10µM) and H2O2 (100µM) strongly relaxed U46619 pre-constricted intact MA (LY83583: 86 ± 6% relaxation at 15min, n=10; H2O2: 94 ± 1% relaxation at 5min, n=2), and the former was partially prevented by combined pre-incubation of SOD and catalase (40 ± 14% relaxation at 15min, n=9, P<0.01 vs. LY83583 alone). In contrast, when MA were pre-constricted with 30 mM KCl, LY83583 caused further sustained constriction and no relaxation (119 ± 15% enhancement at 5 min, n=4, P<0.05). In alpha-toxin permeabilized MA (pre-constricted with 100nM U46619 at pCa 6.4), LY83583 caused modest but concentration-dependent relaxation (10µM: 24 ± 3% relaxation, n=15, P<0.01 vs. vehicle; 30µM: 60 ± 5% relaxation, n=12, P<0.01 vs. DMSO vehicle). LY83583-induced relaxation was prevented by the Rho-kinase inhibitor Y27632 (10µM, n=6, LY83583 relaxation not significant vs. vehicle), but was insensitive to blockade by either the antioxidant tempol (LY83583: 37 ± 6% relaxation, n=6, P=0.05 vs. vehicle) or SOD (LY83583: 36 ± 3% relaxation, n=7, P<0.05 vs. vehicle). The underlying pCa 6.4 constriction was insensitive to LY83583 or vehicle. H2O2 however, and in contrast to its effect in non-permeabilized MA, caused constriction at pCa 6.4 both in the absence (43 ± 7% enhancement at 15min, n=5, P<0.01 vs. time control) and presence of U46619 (17 ± 3% enhancement at 15min, n=4, P<0.001 vs. time controls). In summary, LY83583 either relaxes or constricts MA, depending on the nature of the pre-constriction. LY83583-generated superoxide more strongly relaxes intact compared to permeabilized U46619 pre-constricted MA, whereas exogenous H2O2 relaxes intact but constricts alpha-toxin permeabilized MA. ROS therefore contribute to multiple constriction/relaxation pathways in MA. LY83583-induced Rho-kinase dependent Ca2+-desensitization in MA does not appear to be mediated via ROS.