Urotensin I, a vasoactive peptide isolated from fish urophyses, has previously been demonstrated to cause specific mesenteric dilation in the dog. This mechanism of action should limit its maximal cardiovascular actions: no additional cardiovascular effects should ensure once maximal mesenteric vasodilation is achieved. Provided that the mesenteric vasodilatation does not result in decreased flow to other organs, the agent may, therefore, offer a theoretical advantage as an afterload reducing agent. In pentobarbital anesthetized dogs which were in heart failure as a result of chronic aortico -- left atrial shunt, the effects of urotensin I on cardiovascular hemodynamics were compared with the effects of a nonspecific vasodilator, sodium nitroprusside. At equidepressor doses (approximately 15% decrease in mean arterial pressure; sodium nitroprusside, 2 micrograms . kg-1 . min-1 i.v.; urotensin I, 10 mU . kg-1 . min-1 i.v.), both agents produced comparable falls in total peripheral resistance, left ventricular end diastolic pressure, and pulmonary capillary wedge pressure. Forward cardiac output was increased by both substances, although the increases were not statistically significant. Shunt flow, estimated by echocardiography, was reduced by both. In spite of the marked similarity in hemodynamic effects in this model, the two agents affected regional blood flows differently. Sodium nitroprusside did not significantly alter regional flows measured by radioactive microspheres, although calculated splanchnic, skin, and myocardial vascular resistances were reduced. In contrast, urotensin I, as in the normal dog, greatly increased mesenteric blood flow; this redistribution of cardiac output did not, however, result in underperfusion of other vital organs. These data suggest that urotensin I may be a useful agent in the reduction of afterload in heart failure, particularly since the unique mechanism of action appears to minimize the potential for adverse effects due to excessive dosage.