Eight men and eight women trained 3 days/wk for 16 wk by doing attempted ballistic unilateral ankle dorsiflexions against resistance that either rendered the resultant contractions isometric (one limb) or allowed a relatively high-velocity (5.23 rad/s on an isokinetic dynamometer) movement (other limb). Training sessions consisted of five sets of 10 contractions of each type. Training produced the same high-velocity-specific training response in both limbs (P < 0.001). Peak torque increased most at 5.23 rad/s (38%) in comparison to lower velocities (0, 0.26, 0.52, 1.04, 1.55, 3.02, and 4.19 rad/s). Both limbs also showed similar increases in voluntary isometric rate of torque development (26%) and relaxation (47%) and in evoked tetanus rate of torque development (14%). A similar decrease in evoked twitch time to peak torque (6%) and half-relaxation time (11%) was also observed. Thus, all of these training responses, previously associated specifically with high-velocity resistance training, were produced by a training regimen that prevented an actual rapid movement through a range of movement. The results suggest that the principal stimuli for the high-velocity training response are the repeated attempts to perform ballistic contractions and the high rate of force development of the ensuing contraction. The type of muscle action (isometric or concentric) appears to be of lesser importance.