Forty-one fused hips in thirty-eight patients were converted to total hip replacement. The average length of follow-up was seven years. The predominant indications for conversion were progressively disabling pain in the low back or the hip, or both; loss of function due to immobility or malposition of the hip; and progressive pain and instability of the knee (usually ipsilateral). The postoperative arc of flexion averaged 87 degrees. Limb-length discrepancies improved an average of 2.5 centimeters. Postoperative function of the abductor muscles depended on the preoperative quality of those muscles and on the accuracy of the biomechanical restoration. Postoperative strength of the muscles of the hip improved for two years or more in most patients. There was complete or major relief of pain, improved mobility of the hip, and decreased dependence on supports for walking. There were nine failures: four because of sepsis, four because of loosening of the femoral component, and one because of malposition of the acetabular component. The failures were predominantly in patients who were fifty years old or less at the time of arthroplasty, patients who had had two or more previous operations, and patients who had had an injury to the hip. The quality of the results approached that after primary hip arthroplasty in older patients who have not had multiple previous operations on the hip. Survivorship analysis of the spontaneously fused hips that were treated with conventional hip replacement predicted a probability of survival of the implant of 96 per cent at thirteen years postoperatively (p = 0.048).