We have used the ¹⁸O exchange technique to measure the CO₂ permeability, P_CO2 , of human red blood cells. In normal human red cells we find a P_CO2 of 0.15 ± 0.08 cm/s (S.D.; n=85) at 37°C. To study the possible role of aquaporin-1 (AQP1) for CO₂ permeation, we investigated red cells from patients lacking AQP1 (Colton null). We found a P_CO2 of 0.065 ± 0.028 (n=37), i.e. a reduction from 0.15 cm/s by 0.085 cm/s or by 60%. To investigate a possible contribution of Rh proteins, which have been shown to constitute a pathway for gaseous NH₃, we determined P_CO2 in red cells from patients deficient in Rh proteins (Rh null). In these red cells, we find a P_CO2 of 0.066 ± 0.022 cm/s (n=27), i.e. a reduction from the control P_CO2 of 0.14 cm/s by 0.074 cm/s. Thus, we predict that P_CO2 in the absence of both AQP1 and Rh will be lower by 0.16 cm/s, i.e. close to 0. The AQP1 pathway is found to be inhibitable by pCMBS, and both CO₂ pathways are efficiently inhibited by DIDS. AQP1 deficient red cells in the presence of DIDS show a P_CO2 of 0.012 cm/s, which likely constitutes a "basal" permeability that prevails in the absence of protein CO₂ channels and may be due to the lipids of the membrane. We conclude that both AQP1 and Rh act as CO₂ channels and contribute about equally to the protein-mediated ten-fold increase in CO₂ permeability over the "basal" membrane permeability that presumably is due to the lipid phase.