The minute-by-minute net water movement (J(w)) in the rat jejunum was studied in relation to the diffusive water (P(w)) and mannitol (P(s)) permeabilities with the following results. (a) J(w) was a linear function of the applied hydrostatic and osmotic transepithelial gradients (hydrostatic permeability coefficient, P(hydr) = 0.052 +/- 0.011 cm s-1; osmotic permeability coefficient, P(osm) = 0.0069 +/- 0.0014 cm s-1. (b) A fraction of this absorptive J(w) (transport-associated J(w), J(wt) = 0.086 +/- 0.024-mu-l min-1 cm-2 was independent of the presence of any osmotic, hydrostatic or chemical gradient. (c) In the absence of Na+, J(wt) was not significantly different from zero and there was an increase in P(hydr) but no change in P(osm). (d) In the presence of a hydrostatic gradient (10 cm H2O, mucosal side), acidification of the medium (95% CO2 bubbling, pH 6.2) simultaneously and reversibly increased J(w) and decreased P(w). (e) When an osmotic gradient was present (40 mM polyethyleneglycol on the serosal side) a net increase in J(w) was observed. CO2 bubbling in these conditions reversibly reduced J(w) while increasing P(s). (f) These effects were not observed when the serosal or mucosal pH was reduced in the presence of a nonpermeant buffer (HEPES/TRIS; MES/TRIS). If we accept that P(s) is a good marker of paracellular movements and that P(w) mainly reflects transcellular water movements, we may conclude that acidification of the medium, in the presence of bicarbonate, modifies both paracellular and transcellular routes. The experimental evidence indicates that an increase in proton concentration opens the paracellular pathway and probably has a blocking effect on a transcellular route.