We studied salt and water absorption in isolated rabbit superficial proximal straight tubules perfused and bathed with solutions providing op- positely directed transepithelial anion gradients similar to those which might obtain in vivo. The perfusing solution contained 138.6 mM CI- and 3.8 mM HCO~- (pH 6.6) while the bathing solution contained 113.6 mM CI- and 25 mM HCO~- (pH 7.4); the system was bubbled with 95% O~-5% COs. At 37°C, net volume absorption (J~ nl min -1 mm -1) was 0.32 4- 0.03 (SEM); V,, the transepithelial voltage (millivolts; lumen to bath), was +3.1 4- 0.2. At 21 °C, V, rose to +3.7 -4- 0.1 and J~ fell to 0.13 4- 0.01 (significantly different from zero at P < 0.001); in the presence of 10 .4 M ouabain at 37°C, V, rose to +3.8 4- 0.1 and J~ fell to 0.16 4- 0.01 (P < 0.001 with respect to zero). In paired experiments, the ouabain- and temperature-insensitive moieties of J~ and V, became zero when transepithelial anion concentration gradients were abolished. Titrametric determinations net chloride flux at 21°C or at 37°C with 10 -4 M ouabain showed that chloride was the sole anion in an isotonic absorbate. And, combined electrical and tracer flux data indicated that the tubular epithelium was approximately 18 times more permeable to C1- than to HCO~-. We interpret these results to indicate that, in these tubules, NaC1 absorption depends in part on transepithelial anion concentration gradients similar to those generated in vivo and in vitro by active Na + absorp- tion associated with absorption of anions other than chloride. A quantitative analysis of passive solute and solvent flows in lateral intercellular spaces indicated that fluid absorption occurred across junctional complexes when the osmolality of the lateral intercellular spaces was equal to or slightly less than that of the perfusing and bathing solutions; the driving force for volume flow under these conditions depended on the fact that ~,~coa exceeded ~cl •