In corneal endothelium there is evidence for basolateral entry of bicarbonate into corneal endothelial cells via NBC proteins, and for net bicarbonate flux from the basolateral to the apical side. However, how bicarbonate exits the cells through the apical membrane is unclear. We determined here that cultured corneal endothelial cells transport bicarbonate similarly to fresh tissue. In addition, Cl^-_channel inhibitors decreased fluid transport by at most 16%, and inhibition of membrane-bound carbonic anhydrase IV by benzolamide or dextran-bound sulfonamide decreased fluid transport by at most 29%. Therefore, more than half of the fluid transport cannot be accounted either by anion transport through apical Cl^- channels, or by CO₂ diffusion across the apical membrane, or by a combination of these two mechanisms. However, immunocytochemistry using optical sectioning by confocal microscopy and cryosections revealed the presence of NBC transporters in both the basolateral and apical cell membranes of cultured bovine corneal endothelial cells and freshly isolated rabbit endothelia. This newly detected presence of an apical NBC transporter is consistent with its being the missing mechanism sought. We discuss discrepancies with other reports, and provide a model that accounts for the experimental observations by assuming different stoichiometries of the NBC transport proteins at the basolateral and apical sides of the cells. Such functional differences might arise from either the expression of different isoforms or from regulatory factors affecting the stoichiometry of a single isoform.