A single series of exercises, as Clarke and Mikanagi et al. used in their study, is insufficient for an accurate assessment of muscle fatigue, and our method of using 3 sets of exercise may be more effective for further studies on the characteristics of muscle fatigue and recovery. Our reproducibility test showed that the mean coefficient of variations of individual PT/BW ratio values was 9.6% for the extensor muscles and 17.1% for the flexors, with more scattering in the Flexor muscle values. The mean coefficient of the fatigue rate was 5.1% for the extensor muscles and 9.2% for the flexors, and these coefficients were about 3% lower than the mena coefficient of the SDI, which suggested that our method showed better reproducibility and should be used for examining muscular fatigue. Our result also suggested that the physical condition of the subject on the day of study affected the scattering of muscle strength data, but that the fatigue rate did not change much and was less dependent on minor variations in physical condition