Seven healthy female subjects performed brief (less than 10 s) periods of maximal exercise on a constant-velocity cycle ergometer, over the functional range of pedaling velocities, and an isometric contraction with each leg. There was an inverse relationship between peak torque and pedal crank velocity in all subjects; isometric torque was (mean +/- SE) 19.8 +/- 8.3% greater than the torque recorded at the slowest velocity of 11 rpm. The torque-velocity relationship was described best by a single exponential equation: y = 189.6 X e-0.0834x, where y is peak torque in Newton . meters and x is crank velocity in revolutions per minute. Peak power was a parabolic function of crank velocity; the data were fitted suitably by a second-order polynomial equation: y = -0.0589x2 + 14.504x + 47.092, where y is peak power in watts and x is crank velocity in revolutions per minute. Maximal peak power occurred at crank velocities ranging from 120 to 160 rpm, when the torque was 0.36 +/- 0.06 of the maximal isometric tension. These results demonstrate the importance of recording velocity in measurements of dynamic maximal power.