Hydrogen sulphide (H₂S) has recently joined the ranks of NO and CO as the third gasotransmitter. Although H₂S dilates systemic arteries (Zhao et al., 2001), pulmonary arteries respond to H₂S with a complex constriction-relaxation-constriction (Olson et al., 2006), the mechanism of which we investigated in this study using the H₂S donor sodium hydrogen sulphide (NaHS: 300 µM, [H₂S]≈100 µM). Rat intrapulmonary arteries (IPA) were mounted on a myograph containing physiological saline solution (PSS) and gassed with air/5% CO2 (pH 7.4, 37C), and in some cases were permeablised with α-toxin at 25C, and clamped to pCa 7.1. Intracellular calcium was measured using fura-PE3. All drugs were added before H₂S. IPA slightly preconstricted with U46619 exhibited a triphasic constriction-relaxation-constriction response to H₂S, which persisted in the presence of the α-adrenoceptor blocker phentolamine (10 μM, n=4) or the purinergic receptor blocker suramin (300 μM, n=3). The initial transient constriction (phase 1) and the subsequent sustained constriction (phase 2) were retained in endothelium-denuded IPA (n=4); however the intervening relaxation was suppressed by removal of the endothelium or treatment with L-NAME (300 µM; n=4&7). Phase 1 but not phase 2 was decreased by thapsigargin (1 µM; n=5; P<0.05) or ryanodine (100 µM, n=4; P<0.01). Phase 1 was unaffected by the Ca²⁺ channel antagonists nifedipine (3 µM, n=2) or diltiazem (10 µM, n=3). In IPA bathed in Ca²⁺-free PSS (100 μM EGTA), H₂S elicited a rise in [Ca²⁺]i that was suppressed by cyclopiazonic acid (30 µM), n=2). In tension experiments, the Rho kinase inhibitor Y-27632 (1 µM, n=5; P<0.05) but not the broad spectrum protein kinase C inhibitor Gö6983 (3 µM, n=3) inhibited phase 2, while in permeablised arteries, H₂S caused a contraction additive to that induced by pCa 7.1 (100 µM NaHS, n=4; P<0.05; 300 µM NaHS, n=6; P<0.05). Notably, the increase in tension induced by H₂S in IPA under slight preconstriction was not associated with a concomitant increase in [Ca²⁺]i. We conclude that the constriction-relaxation-constriction response elicited by H₂S in rat IPA is due to 1) an initial endothelium-independent contraction associated with Ca²⁺ release from ryanodine sensitive stores, 2) an endothelium and NO-dependent relaxation, and 3) finally a Rho kinase-dependent but endothelium-independent contraction that is not associated with an increase in [Ca²⁺]i. In view of some similarities to the response to hypoxia, we speculate that H₂S elicits its response in IPA in part via inhibition of cytochrome oxidase, this mimicking hypoxia.