Total hip replacement was done in thirty-one patients who had had spontaneous fusion of the hip and previous surgical attempts at arthrodesis of the hip. Follow-up was from eighteen months to twelve years, the average being four years and two months. Eighteen hips had had spontaneous ankylosis and thirteen hips had had surgical ankylosis. The indications for the total hip replacement were loss of function due to immobility, and pain in the hip and ipsilateral knee and low back, and malposition of the ankylosed hip. Satisfactory result was obtained in twenty-five of thirty-one patients (eighty-one %). Postoperatively the average range of flexion was 84 degrees, and the range of abduction-adduction averaged 51 degrees, the range of rotation averaged 51 degrees, and the shorter the duration of ankylosis was, the more the range of motion was achieved. Whether to use cement or not has no difference at present. Postoperative limb discrepancies improved an average of 1.5 centimeters. The duration of ankylosis had not effect on result of the total hip arthroplasty. Postoperatively the strength of the abductor was gradually increased. All cases who recieved the total hip arthroplasty were satisfactory due to improved motion of the hip joint and improved leg length discrepancy and disappearance of pain. One hip was complicated by deep infection, and four hips complicated by aseptic loosening, and two hips complicated by an ectopic bone formation, and one hip complicated by dislocation, and three hips complicated by transient femoral nerve stretching.