Vasopressin (VP) was administered for 1 h intravenously to hydropenic, anesthetized dogs in doses of 1.0-1.25 mU/kg per min. In 14 experiments, sodium excretion (UNA V) increased from a mean of 13 +/- 5 to a peak of 96 +/- 21 mueq/min 40 min after beginning infusion (P less than .001). Urine flow and potassium excretion increased from 0.18 +/-.04 ml/min and 20 +/- 2 meuq/min to peak values of 0.6 +/- .08 ml/min and 61 +/- 9 mueq/min, respectively (P less than .001), with no significant increase in glomerular filtration rate. No significant changes in UNA V occurred in eight sham control experiments of in six experiments in which VP was given at 75 muU/kf per min. To test the hypothesis that VP might be natriuretic indirectly by releasing a natriuretic substance, plasms ultrafiltrates were tested for toad bladder antinatriferic activity(AA). During steady-state control, AA was -10 +/- 3%. Thirty and sixty minutes after beginning VP, AA increased to -24 +/- 3% (P less than .05) and -26 +/- 2% (P less than .001), respectiviely. No significant change in plasma AA occurred in either sham controls or in animals given the subnatriuretic VP dose. Incubation of plasma with 1,000 muU/ml VP caused no increase in AA. The data show that VP natriuresis is accompanied by an increase in plasms AA. The results suggest that vasopressin natriuresis in hydropenic dogs at least in part to the release of a humoral inhibitor of renal tubular sodium transport.