Renal function has been studied by the clearance (cl.) method during hypotonic polyuria--four 15-min cl. periods--and successive antidiuresis--two 60-min cl. periods (A1, A2)--induced by lysine-8-vasopressin (LVP), 5 mU in bolus followed by infusion at a rate of 0.04 mU/min. The endogenous creatinine cl. (Cc) and the osmotic cls. (Cosm, CH2O) were determined by the usual methods as well as the absolute and fractional urinary excretions of water, sodium, chloride and potassium. The urinary concentrations of PGE2, 6-keto-PGF1 alpha and TxB2 were determined by the RIA method. This study protocol has been applied to 28 healthy women either in normal potassium balance (N, n = 14) or after potassium depletion (KD) induced by low potassium dietary intake (less than or equal to 10 meq/d) plus natriuretic treatment according to two different time patterns: two KD groups were obtained with potassium cumulative deficit of 160 +/- 43 (D2, n = 8) and 198 +/- 22 meq (D3, n = 6). The early % effects of LVP, i.e. (A1-P)% of P (mean polyuria), were significantly different only in D3 as compared to N. Precisely, the LVP-effect to reduce Cc was blunted; moreover a LVP-effect to reduce renal sodium and chloride fractional excretions and a tendentiously enhanced LVP-effect to reduce water fractional excretion were observed. These tubular effects are likely related to the inhibited renal synthesis of prostanoids in the D3 group.