During a twelve-month period, 23 per cent of a series of knees undergoing total replacement were manipulated two weeks postoperatively to increase flexion. While manipulation achieved an immediate increase in flexion from a mean of 71 degrees to a mean of 108 degree, by one week after manipulation the mean flexion was reduced to 88 degrees. By one year postoperatively, the manipulated knees were found to have a range of motion similar to that of their non-manipulated counterparts. Such factors as preoperative flexion and diagnosis appear to be the major determinants of ultimate flexion, and they seem to offset the temporarily increased flexion afforded by manipulation. The primary reason for manipulation is to facilitate the postoperative rehabilitation program for patients with painful, limited motion of the knee.