Studies in man produce conflicting evidence of the role of the Frank-Starling mechanism in increasing cardiac output during exercise, though animal studies indicate that it may help to improve cardiac performance during severe exertion. Twelve healthy volunteers (mean age 35.8 i 2.8 years) performed graded exercise to exhaustion on a bicycle ergometer in the semisupine position for 8.9 i 0.9 minutes (maximum work load 900 kg-m/min). Echocardiographic recordings of left ventricular dimensions were obtained continuously and end-expiratory tracings digitized. Heart rate increased from 64 ± 3 to 152 ± 4 beats/min. At peak exercise, end-diastolic diameter increased from 4.52 ± 0.20 to 5.24 ± 0.17 cm sec-1 at peak exercise (p>0.0005). Maximum rates of change in diameter in systole and diastole, and normalized maximum diastolic rate of change all increased progressively and significantly throughout the exercise period. These results suggest that severe semisupine exertion causes an increase in left ventricular end-diastolic diameter, stroke dimension and percent change in diameter, but no change in end-systolic diameter measured at end-expiration. Increases in indices of left ventricular fiber shortening and rates of lengthening appear earlier in exercise than does an increase in end-diastolic fiber length, suggesting that during lower levels of exertion cardiac output rises primarily by increases in heart rate. The Frank-Starling effect appears, under the conditions of this study, to be reserved for augmenting cardiac performance during severe semisupine exertion.