Because the results of treatment of osteochondritis dissecans of the femoral condyles are still not consistently satisfactory, we reviewed the cases of eighty-three patients (ninety-five knees) with osteochondritis dissecans who were followed for two to thirty-one years to identify factors that may influence treatment and long-term prognosis. Of the eighty-three patients, sixteen had only non-surgical treatment, sixty-five had only surgical treatment, and two had non-surgical treatment of one knee and surgical treatment of the other. Of the twenty-two knees (eighteen patients) that were treated non-surgically, fifteen were treated before and seven were treated after distal femoral epiphyseal closure. Of the seventy-three knees (in sixty-seven patients) that were treated surgically because of persisting symptoms, twenty-three were treated before epiphyseal closure and fifty, after closure. At follow-up, each knee was given a rating of excellent, good, fair, poor, or failure, and a corresponding point score. Seventy-seven per cent of the knees in the surgical group and 82 per cent of those in the conservatively treated group were rated either excellent or good. The average scores in both groups were higher in knees in which the osteochondritic defect was small and was treated before epiphyseal closure, and in knees in which the fragment healed as compared with the ones from which the fragment was removed. We found that knees with osteochondritis dissecans of the femoral condyles that had no other abnormal physical findings or functional disability responded well to conservative treatment before epiphyseal closure.(ABSTRACT TRUNCATED AT 250 WORDS)