1 In human pulmonary vascular preparations, precontracted arteries were more sensitive to the relaxant effect of acetylcholine (ACh) than veins (pD 2 values: 7.2570.08 (n ¼ 23) and 5.9270.09 (n ¼ 25), respectively). Therefore, the role of prostacyclin (PGI 2 ) was explored to examine whether this mediator may be responsible for the difference in relaxation. 2 In the presence of the cyclooxygenase (COX) inhibitor, indomethacin (INDO), the ACh relaxations were reduced in arteries but not in veins. On the contrary, an inhibitor (L-NOARG) of the nitric oxide synthase blocked preferentially the relaxation in veins. 3 A greater release of 6-keto-PGF 1a , the stable metabolite of PGI 2 , was observed in arterial preparations than in venous preparations when stimulated with either ACh or arachidonic acid (AA). 4 Exogenous PGI 2 produced a reduced relaxant effect in the precontracted vein when compared with the artery. In the presence of the EP 1 -receptor antagonist AH6809, the PGI 2 relaxation of veins was similar to arteries. 5 In veins, AA (0.1 mM) produced a biphasic response, namely, a contraction peak (0.4-0.5 g) followed by a relaxation. These contractions in venous preparations were abolished either in the absence of endothelium or in the presence of INDO or an EP 1 -receptor antagonist (AH6809, SC19220). In the arterial preparations AA induced only relaxations. 6 In both vascular preparations, COX-1 but not the COX-2 protein was detected in microsomal preparations derived from homogenized tissues or freshly isolated endothelial cells. 7 The differential vasorelaxations induced by ACh may be explained, in part, by a more pronounced production and release of PGI 2 in human pulmonary arteries than in the veins. In addition, while PGI 2 induced relaxation by activation of IP-receptors in both types of vessels, a PGI 2 constrictor effect was responsible for masking the relaxation in the veins by activation of the EP 1 -receptor.