We have used rat inner medullary collecting ducts (IMCD) perfused "in vitro" to study the effect of vasopressin (VP) on the unidirectional Na+ flux (in nmol.cm-2.s-1). We found that, at a high perfusion rate in the basal state, 24Na lumen-to-bath flux (Jl----b) was greater than the bath-to-lumen flux (Jb----l) (4.88 +/- 0.15 vs. 2.57 +/- 0.21), resulting in a significant net flux (Jnet) (P less than 0.001). Addition of 10 microU/ml of lysine vasopressin (LVP) to the bath produced a stable increase in Jl----b to 6.66 +/- 0.35 (P less than 0.001) without significant effect on Jb----l. Measuring directly the net flux absorption at lower perfusion rate (8 nl/min), we observed that LVP (10 microU/ml) produced a reversible stimulation on Jnet from 1.39 +/- 0.14 to 2.79 +/- 0.23 (P less than 0.01). The transtubular potential difference (PD) measured in the middle and final third of IMCD showed a small but significant PD (0.30 +/- 0.02 mV lumen positive) that increased significantly to 0.60 +/- 0.04 mV in the presence of LVP. However, dibutyryl adenosine 3',5'-cyclic monophosphate (DBcAMP, 10(-4) M) added to the bath fluid did not change the JNa+l----b, nor was 1-desamino-8-D-arginine vasopressin (50 microU/ml), a specific V2-agonist, able to increase the Na+. We also demonstrated that JNa+l----b stimulated by LVP from 4.70 +/- 0.08 to 6.33 +/- 0.26 (P less than 0.01) was completely and reversibly inhibited by V1-antagonist, d(CH2)Tyr(Me)AVP, to 4.79 +/- 0.05. On the other hand, the absence of Ca2+ in the bath or the addition of amiloride to the lumen fluid or ouabain to the bath fluid completely inhibited AVP-stimulated JNa+l----b. Therefore, AVP and LVP increase Na+ absorption in the rat IMCD by increasing the Na+ outflux, probably generated by an increase of luminal membrane Na+ permeability modulated by extracellular Ca2+ and mediated through V1-receptors and independent of cAMP cascade.