We investigated the capacity for torque development and muscle activation at the onset of fast voluntary isometric knee extensions at 30, 60, and 90 degrees knee angle. Experiments were performed in subjects (n = 7) who had high levels (>90%) of activation at the plateau of maximal voluntary contractions. During maximal electrical nerve stimulation (8 pulses at 300 Hz), the maximal rate of torque development (MRTD) and torque time integral over the first 40 ms (TTI40) changed in proportion with torque at the different knee angles (highest values at 60 degrees ). At each knee angle, voluntary MRTD and stimulated MRTD were similar (P < 0.05), but time to voluntary MRTD was significantly longer. Voluntary TTI40 was independent (P > 0.05) of knee angle and on average (all subjects and angles) only 40% of stimulated TTI40. However, among subjects, the averaged (across knee angles) values ranged from 10.3 +/- 3.1 to 83.3 +/- 3.2% and were positively related (r2 = 0.75, P < 0.05) to the knee-extensor surface EMG at the start of torque development. It was concluded that, although all subjects had high levels of voluntary activation at the plateau of maximal voluntary contraction, among subjects and independent of knee angle, the capacity for fast muscle activation varied substantially. Moreover, in all subjects, torque developed considerably faster during maximal electrical stimulation than during maximal voluntary effort. At different knee angles, stimulated MRTD and TTI40 changed in proportion with stimulated torque, but voluntary MRTD and TTI40 changed less than maximal voluntary torque.