The aim of this study was to assess the effects in humans of early (2 weeks) and delayed (6 weeks) isokinetic strength training in the recovery of muscle strength following an arthroscopic partial meniscectomy. The peak torque developed in the quadriceps and hamstrings and the torque developed at a knee angle of 1.05 rad were evaluated in 16 subjects, pre-operatively (pre-op), and 2, 6, and 10 weeks post-operatively (post-op), on an isokinetic device at four different velocities (1.05, 2.09, 3.14, and 4.19 rad.s-1). The fatigue characteristics of the muscles were evaluated by having the subject perform 15 maximal contractions at 3.14 rad.s-1. Training was done on the same device (three times a week for 1-2 months), beginning either 2 or 6 weeks post-op. A repeated measures analysis of variance demonstrated a time effect but no differences between groups and no interactions. Torques developed by the knee flexors and extensors were significantly smaller 2 weeks post-op than pre-op, at all velocities tested. Torques developed in the quadriceps recovered to their pre-op values by 6 weeks, and further gained significantly in strength from 6 to 10 weeks. Quadriceps torques remained weaker than the contralateral side at 10 weeks. Hamstrings torques were either higher or similar to pre-op values by 6 weeks, and demonstrated increases from 6 to 10 weeks post-op at 1.05 and 4.19 rad.s-1 only. Total work and average power developed by the quadriceps and hamstrings during the fatigue protocol changed with time in a similar manner to torque.(ABSTRACT TRUNCATED AT 250 WORDS)