Arginine vasopressin (AVP) was infused into the renal artery of seven uninephrectomized conscious dogs at successive rates of 0.09, 0.36, and 1.46 ng X kg-1 X min-1 for 18, 9, and 5 days, respectively; subsequently, the nonpressor analogue of AVP, 1-desamino-8-D-arginine vasopressin (DDAVP), was infused intrarenally for 7 additional days. The lowest infusion rate of AVP produced a high concentration of AVP in the renal circulation (maximal antidiuresis) with only a relatively moderate increase in peripheral plasma AVP concentration. For comparison, the effects of a comparable increase in peripheral plasma AVP concentration during intravenous infusion at this same rate were observed in an additional group of dogs. Acutely, when this low dose of AVP was infused either intrarenally or intravenously, there was marked antidiuresis, but there were no significant changes in mean arterial pressure (MAP), renal hemodynamics, or urinary electrolyte excretion. Chronically, in both groups of animals, cumulative water balance was positive, and glomerular filtration rate, effective renal plasma flow, and MAP (16-18 mmHg) all increased; plasma renin activity decreased. Similar changes were observed during DDAVP infusion. When elevated peripheral plasma levels of AVP were achieved during the higher infusion rates of AVP, natriuresis and diuresis occurred, but, otherwise there was little change in the above variables, including MAP. Thus the hydrosmotic effects of AVP appear to account for its moderate hypertensive activity. Further, the failure of AVP to produce prominent hypertension, even when pronounced systemic vasoconstrictor effects are manifested, may be a result of its inability to promote significant renal vasoconstriction and antinatriuresis.