Sixty-six total knee replacements in 50 morbidly obese women with osteoarthritis were performed. The results in these patients were compared with 64 knee replacements in 50 nonobese women and with 50 obese women who did not have knee replacement operations. All obese patients were advised by a physician to lose weight and were counseled by a dietician. Five obese patients with six replaced knees and four nonobese patients were lost prior to follow-up study. Follow-up examinations were performed after 24 to 60 months, with mean of 33 months, in the remaining patients. No significant sustained weight loss occurred in the obese patients who had knee replacement operations or in obese patients who did not have surgical treatment. Eighty-six per cent of nonobese patients had excellent or good results from knee replacement surgical treatment. Fifty-seven per cent of morbidity obese patients had excellent or good results. There were two infections and two revision operations in morbidly obese patients and no infections and one revision in nonobese patients. Total knee replacement is safe and reasonably effective in the morbidly obese. However, the results are not comparable with those achieved in nonobese patients, and knee replacement should not be expected to facilitate weight loss.