We investigated whether correction for body mass is appropriate when using variables derived from gas exchange measurements in the assessment of chronic heart failure. Patients with heart failure were studied either on a bicycle ergometer (22 subjects) or on a treadmill (28 subjects). For bicycle exercise testing either peak V<sub>O2</sub> or peak V<sub>CO2</sub> uncorrected for body mass reflected exercise capacity. However, for treadmill exercise, body mass correction was necessary to relate exercise capacity to peak V<sub>O2</sub> or peak V<sub>CO2</sub>. Mass correction is therefore appropriate for treadmill but not for bicycle exercise.