The use of anaerobic threshold in assessment of aerobic capacity was evaluated in 34 normal subjects and 47 patients with various kinds of chronic heart disease. Anaerobic threshold was determined as the oxygen consumption (VO2) at which a linear relationship between pulmonary ventilation (VE) and VO2 was lost during progressive treadmill exercise. Anaerobic threshold determined in this manner was validated with that determined by blood lactate measurements in eight normal subjects and nine cardiac patients (r = .962, p less than .001). Thereafter, anaerobic threshold was determined only by respiratory measurements. In symptom-limited, maximal exercise, anaerobic threshold was reached well before maximal effort and corresponded to 70% of maximal VO2 both in normal subjects and cardiac patients. Anaerobic threshold decreased as age progressed in normal subjects (r = - .70, p less than .001). Anaerobic threshold in cardiac patients was lower than that in the normal subjects and decreased progressively as New York Heart Association functional classification advanced (normal, 32.95 +/- 6.17 ml/min/kg; class I, 22.78 +/- 3.74; class II, 16.99 +/- 3.66; class III, 12.97 +/- 2.76; p less than .01 between each group other than between class II and class III). Anaerobic threshold in cardiac patients correlated poorly with other objective indices, e.g., cardiomegaly (r = -.54, p less than .001) and rise in pulmonary wedge pressure (r = -.64, p less than .001). At anaerobic threshold, cardiac patients subjectively graded the work load as light (13%), light-to-moderate (27%), moderate (30%), and moderate-to-heavy (28%). Thus determination of anaerobic threshold by respiratory measurements is a safe, accurate, and objective method to measure aerobic capacity in cardiac patients and in normal subjects.