A series of experiments was undertaken in a total of 95 adult subjects, of whom 25 had coronary artery disease, to evaluate the extrapolation (Defares) CO2 rebreathing method vs. the equilibrium (Collier) method for estimating mixed venous CO2 tension. Collier values were corrected for the downstream effect, whereas Defares values were uncorrected. Although the methods gave similar mean values in separate series with normal subjects walking on a treadmill set at 123 W for 3 min, Collier values had a coefficient of variation (CV) of 2.5% in duplicate determinations and Defares values had a CV of 4.5%. In paired comparisons Defares values averaged either higher or lower than Collier values depending on variations in technique and the analysis of Defares tracings. Collier values were essentially unaffected by the duration of the rebreathing period (10 vs. 15 s). The Collier technique appears to be superior for the exercise evaluation of cardiac output in healthy and diseased subjects when the goal is obtain values at 82 or 123 W in single test sessions.