To examine the role of vasopressin V 1 and V 2 receptors, nitric oxide and prostanoids in the coronary vascular effects of [Arg 8]vasopressin, coronary blood flow was measured with an electromagnetic flow transducer placed around the left circumflex (23 goats) or anterior descending (11 goats) coronary artery and vasopressin (0.03–1 μg) was intracoronarily injected in 34 anesthetized, open-chest goats. Basal mean values for coronary blood flow, mean systemic arterial pressure and heart rate, were 34±2.38 ml/min, 89±3.34 mmHg and 80±3.06 beats/min, respectively. Vasopressin produced dose-dependent decreases in coronary blood flow and the maximal reduction of this flow, attained with 1 μg of vasopressin, was 14±1.49 ml/min (42±2.64% of basal flow) ( P<0.01). Desmopressin (0.03–1 μg; 8 goats) did not affect significantly coronary blood flow. The intracoronary infusion of the antagonist for vasopressin V 1 receptors d(CH 2) 5Tyr (Me) arginine vasopressin (2 μg/min per kg, 6 animals) significantly diminished the effects of vasopressin on coronary blood flow (the effects of 1 μg of vasopressin were reduced by 28%, P<0.05). The mixed antagonist for vasopressin V 1 and V 2 receptors desGly-d(CH 2) 5- d-Tyr(Et)Val arginine vasopressin (0.2, 0.7 and 2 μg/min per kg, 9 animals) decreased in a dose-dependent manner the effects of vasopressin on coronary blood flow (the effects of 1 μg of vasopressin were decreased by 61% with 2 μg/min per kg, P<0.01). Intracoronary infusion of saline (vehicle, 3 goats) did not change the effects of vasopressin on coronary blood flow. Intravenous administration of the inhibitor of nitric oxide synthesis N w-nitro- l-arginine methyl ester ( l-NAME, 47 mg/kg, 9 animals) decreased resting coronary blood flow by 10% ( P<0.01) and augmented mean systemic arterial pressure by 20% ( P<0.01), without changing heart rate. During this treatment the reduction in coronary blood flow produced by vasopressin was higher than under control (the effects of 1 μg of vasopressin were increased by 28%, P<0.01). Intravenous administration of the inhibitor of cyclooxygenase, meclofenamate (5 mg/kg, 7 animals), neither modified resting coronary blood flow, arterial pressure and heart rate nor the effects of vasopressin on this flow. These data indicate that vasopressin produces marked coronary vasoconstriction and suggest that: (a) it may be mediated by vasopressin V 1 receptors, without involvement of vasopressin V 2 receptors, (b) it is probably inhibited by nitric oxide under normal conditions and (c) it may be not modulated by prostanoids.