A manometric assay has been constructed such that carbonic anhydrase dehydration activity can be determined utilizing intact erythrocytes. It was found that unwashed whole trout blood lacks any dehydration activity quite in contrast to whole rat blood. Removal of the plasma and replacement with physiological saline results in pronounced dehydration activity with inhibition restored by replacing the saline with the original plasma. An inhibitor residing in the plasma (probably a protein but not yet fully characterized) is capable of limiting in vivo dehydration activity. The inhibitor could work at the level of the enzyme per se or by inhibiting erythrocytic HCO3 influx, or both. The lack of erythrocytic dehydration activity presumably would alter CO2 excretion patterns; possible implications are discussed.