1. Duplicate measurements were made of mixed venous P(CO)2 (Pv,(CO)2) by two rebreathing methods during steady-state exercise at three power outputs in seven subjects. One method employed a high initial bag CO2 concentration to obtain equilibrium of CO2 in the lung-bag system before recirculation (equilibrium method); in the other, a low initial bag CO2 concentration was used and a statistical method was applied to alveolar P(CO)2 measurements before recirculation, to obtain the asymptote from the exponential rise in end-tidal P(CO)2 during rebreathing (exponential method). 2. The reproducibility was similar; (SD) of duplicate determinations of Pv,(CO)2 was 0.15 kPa (1.1 mmHg) for the equilibrium method and 0.20 kPa (1.5 mmHg) for the exponential method. Measurements of Pv,(CO)2 by the exponential method were systematically lower than the equilibrium method. When the equilibrium Pv,(CO)2 was corrected for the alveolar-arterial ('downstream') P(CO)2 difference, using published values, Pv,(CO)2 was similar for both methods. 3. As an alveolar to arterial P(CO)2 difference did not appear to exist with the exponential method, it is concluded that the previously described disequilibrium between alveolar and arterial P(CO)2 during rebreathing in exercise is mainly related to prevention of net CO2 movement from the pulmonary capillary blood in the equilibrium method, and is not present when continuous CO2 evolution occurs in the exponential method.