The possibility that small amounts of vasopressin (AVP) reduce water excretion without affecting solute excretion was investigated in conscious dogs. AVP was infused intravenously for 120 min at rates of 2 and 5 pg.min-1.kg body wt-1 during water diuresis elicited by a sustained water load of 2% body wt. During control experiments urine osmolality was constantly approximately 60 mosmol/kgH2O; during AVP infusions it increased by factors of 1.36 (P less than 0.01) and 2.12 (P less than 0.01), respectively, concomitant with 39 +/- 6 and 61 +/- 7% reductions in urine flow. Osmolar and free water clearances decreased significantly. Sodium excretion did not change; changes in potassium excretion during AVP were similar to those of the control series, i.e., a gradual decline. During AVP, 5 pg.min-1. kg-1, creatinine and urea clearances decreased (25 +/- 2 and 31 +/- 7%, respectively, both P less than 0.01). With the assumption of metabolic clearance rates of AVP of 15-40 ml.min-1.kg body wt-1, the increase in plasma AVP during the infusion of 2 pg.min-1.kg body wt-1 was 5-13 X 10(-14) M. It is concluded that small increments in plasma AVP may reduce glomerular filtration rate and that with increasing levels of AVP in plasma 1) reduction of free water clearance, 2) reduction in urea clearance, and 3) natriuresis-kaliuresis occur in that order. Apparently AVP cannot reduce water excretion without changing the rate of excretion of solutes.