Although total knee arthroplasty (TKA) has been studied extensively, objective muscle testing has not been reported. Isokinetic testing of 68 patients with degenerative joint disease scheduled for unilateral TKA revealed that marked muscular deficits in flexion and extension were present preoperatively in the involved knee. Postoperatively, hamstring peak-torque values were able to attain strength levels of the uninvolved knee within the period of seven to 12 months after surgery, whereas the quadriceps mechanism still showed a residual deficit at two years follow-up evaluation. In addition, the ratio of flexion to extension peak torques in the operated knee returned to normal values as the quadriceps mechanism was rehabilitated. A comprehensive evaluation system consisting of Cybex II isokinetic testing, gait mat analysis, and the Hospital for Special Surgery knee rating scale is also presented. Isokinetic testing correlated well with gait analysis. Patients with a nearly balanced quadriceps-to-hamstring ratio walked with a more symmetrical gait pattern. The knee rating scale was less reliable in assessing functional outcome. Functional testing and evaluation at the authors' institution has provided an important source of objective information that allows better planning and evaluation of TKAs. These isokinetic studies enable more critical planning of the rehabilitation program. Hamstring or quadricep exercises may be emphasized as required. The authors conclude that a balanced hamstring to quadriceps mechanism is needed for resumption of normal gait.