These experiments examined the effects of chloride gradients on total CO2 fluxes across the cortical collecting tubule. Isolated perfused rabbit cortical collecting tubules at 38°C were exposed to chloride gradients in either lumen-to-bath or bath-to-lumen direction. Total CO2 flux was determined by microcalorimetry of timed collections of luminal perfusion solutions. Transepithelial electrical potential differences were minimized by addition of 0.1 mM amiloride to the perfusate. When 50 mM of bath chloride was replaced by 50 mM of isethionate or cyclamate, and both perfusate and bath contained 25 mM bicarbonate, total CO2 flux changed from absorption (1.3 ± 1.5 pmol.mm-1.min-1) to secretion (-2.1 ± 1.6 pmol.mm-1.min-1) during the gradient period (P < 0.001). Conversely, when a bath-to-lumen chloride gradient was created (114 mM isethionate replaced 114 mM chloride in the perfusate), total CO2 absorption increased from +8.1 ± 1.9 to +12.9 ± 2.7 pmol.mm-1.min-1 (P < 0.01). In contrast, if methylsulfate replaced bath chloride (50 mM), no change in total CO2 flux was found. If perfusate was initially bicarbonate free and 114 mM chloride gradient from bath to lumen was created by replacing the bath chloride with 114 mM isethionate, bicarbonate secretion decreased from -7.5 ± 2.6 to -5.6 ± 2.1 pmol.mm-1.min-1 (P < 0.05). When the 114 mM bath-to-lumen chloride gradient was then eliminated by replacing bath chloride with isethionate in four experiments, bicarbonate flux returned to control levels, -7.1 ± 2.0 pmol.mm-1.min-1. We conclude that chloride gradients in one direction affect bicarbonate flux in the opposite direction in cortical collecting tubules but cannot distinguish whether the effect is due to electrical coupling or to an anion exchange mechanism.