We used a Cybex II machine as Mikanagi et al. did, but our method was geared toward assessing both muscular fatigue and recovery. Accordingly, our subject performed 50 knee bends at maximum strength. Few of the subject showed maximum torque at the first bend and minimum torque at the last bend. In a study performed by Sargeant et al. in 1981, power output during constant speed cycling reached a peak at 2-3 sec after the start of exercise and declined rapidly. For this reason, we selected the maximum muscle torque among the first 5 bends and the minimum muscle torque among the last 5 bends in order to calculate the fatigue rate with more specific and reliable parameters than those of the SDI. Our subject performed 3 series of bends and observed a 10 min rest between each trail that corresponded to the rest period in a soccer game. This allowed us to examine the pattern of change in the maximum torque and the fatigue rate after rest, so that we could assess recovery as well as fatigue.