Human amnion was mounted, immediately after delivery, as a diaphragm between two lucite chambers and the net transepithelial water movement (Jw) was recorded minute by minute. When Jw was plotted against the applied transepithelial hydrostatic pressure (fetal side positive), in the absence of any other gradient, a linear relationship was observed (Phydr = 0.32 +/- 0.05 cm/s, n = 10). A linear relationship was also found when Jw was measured in the presence of an osmotic gradient, generated by adding (to the maternal side) different concentrations of poly(ethylene glycol) (Mr approximately equal to 3600; reflexion coefficient (sigma) = 1; Posm = 0.015 +/- 0.001 cm/s, n = 10). When sucrose, a paracellular marker, was used as the osmotic probe, the observed sigma was 0.5. Medium acidification in the presence of bicarbonate reduced in the same proportion both the hydrostatic and osmotic permeabilities. The effect was fully reversible, but was not observed when bicarbonate was replaced by Tris. To test the comparative role of transcellular versus paracellular paths, Jw and the [14C]sucrose permeability (Psuc) were simultaneously recorded minute by minute, in the presence of an osmotic or an hydrostatic gradient. In both cases, the percentage reductions in Jw and Psuc induced by medium acidification were similar. Quantification of theoretical and observed values for Jw and Psuc strongly suggests that effects of pH on both the osmotic and hydrostatic flux reflect a modification of the paracellular path.