Intravenous atrial natriuretic peptide (ANP) usually results in splanchnic vasoconstriction in humans or experimental animals that is accompanied by falls in blood pressure and/or cardiac output. To determine direct in vivo effects in the present study, ANP was infused (12 ng. kg-1. min-1) directly into the mesenteric (iMA) and hepatic (iHA) arterial beds of anesthetized dogs, thereby minimizing changes in blood pressure. Over the first 2 min of iMA infusion, rate of change in mesenteric vascular resistance was 19.6 +/- 5.4 mmHg. l-1. min-1/min, reaching a maximum increase in resistance of 22 +/- 4% compared with baseline after approximately 10 min. There was no evidence of vasodilatation at any stage. The mesenteric response was similar whether ANP was infused iMA, iHA, or via the femoral vein (30 ng. kg-1. min-1). In contrast, hepatic vasoconstrictor response to ANP infusion iHA or into the portal vein was only evident after approximately 5 min, reaching a maximum increase in hepatic vascular resistance of 11 +/- 6% after approximately 15 min iHA infusion. When preinfused through the gut vasculature (iMA), ANP increased hepatic vascular resistance earlier and reached similar levels (14 +/- 3%), despite a lower arterial concentration of ANP. It is proposed that a vasoconstrictor agent from the intestinal circulation contributed to ANP-induced splanchnic vasoconstriction.