1 A component of isoprenaline-mediated vasorelaxation in pulmonary arteries is mediated by nitric oxide (NO). We examined the eects of physiological concentrations (4400 mM) of L-arginine on isoprenaline-induced relaxation in rat pulmonary arteries, and following inhibition of L-arginine uptake with L-lysine. In addition, we examined the role of the endothelium, and whether L-arginine aected acetylcholine (ACh)-induced relaxation. 2 Isoprenaline-induced relaxation was potentiated by 400 mM L-arginine in pulmonary arteries; maximum relaxation was increased from 83+4% of initial tone to 94+4% (P50.05). L-lysine (10 mM) not only abolished the potentiation by L-arginine, but suppressed relaxation compared to control (70+4%, P50.05), even in the absence of L-arginine added to the bath. Blockade of NO synthase with 100 mM L-NMMA or removal of the endothelium inhibited isoprenaline-induced relaxation to the same extent as L-lysine, and under these conditions the presence or absence of 400 mM L-arginine made no dierence. L-lysine had no additional eect when applied in combination with L-NMMA. 3 The eect of extracellular L-arginine was concentration dependent, with an apparent EC 50 of *1±7 mM. 4 Relaxation to the membrane permeant cyclic AMP analogue CPT cyclic AMP was also potentiated by L-arginine and inhibited by L-lysine. There was however no dierence in relaxation induced by acetylcholine (ACh) in the presence of L-arginine or L-lysine, and isoprenaline-induced relaxation of mesenteric arteries was unaected by L-arginine or L-lysine. 5 These results strongly suggest that extracellular L-arginine is critically important for development of the NO-and endothelium-dependent component of cyclic AMP-induced vasorelaxation in rat pulmonary arteries, but is not required for ACh-induced relaxation. As the apparent EC 50 for this eect is in the low micromolar range it is likely to be fully activated in vivo, as plasma L-arginine is 4150 mM.