From 1982 to 1984, an alumina ceramic total knee (KC-1 type) replacement without cement was performed in 64 knees of 48 patients. The design of the prosthesis is similar to that of the cementless Mark-II prosthesis (metal on plastic), which originated in our department. Clinical evaluation was made on 52 knees of 40 patients according to a criteria proposed by three universities. Radiographic findings were compared with the the clinical assessment. The mean length of follow-up was 29.9 months, ranging from 10 to 43 months. This prosthesis was inserted in 29 patients (38 knees) with rheumatoid arthritis (RA) and in 11 patients (14 kness) with osteoarthritis (OA). The average age at the time of surgery was 58.4 years, ranging from 43 to 81 years. Clinical knee scores were 36.7 in RA patients and 42.5 in OA preoperatively, and improved up to 74.6 and 82.3, respectively, at the follow-up. The mean range of motion (ROM) was 77.7 degrees in RA and 103.5 degrees in OA preoperatively and 86.3 degrees in RA and 98.2 degrees in OA at the follow-up. The mean flexion contracture of the knees improved from 24 to 5 degrees in RA and from 13.2 to 0.3 degrees in OA. The mean ROM was significantly improved in cases of TKR with bone graft, compared to that in cases of TKR without graft. Pain relief and joint stability were clinically excellent. The average femorotibial angle (FTA), determined by radiographic examination was 176.8 degrees preoperatively, and remained almost unchanged until the time of the follow-up. The average medial angle of the tibial component was 88.7 degrees a few months after the operation, and 88.4 degrees at the time of the follow-up. The average postoperative posterior angle of the tibial component was 84 degrees, the same as at the time of the follow-up. In this series, the postoperative FTA and the medial angle of the tibial component tended to correlate with the pain relief score. In 7 knees (6 patients), the postoperative medial angle of the tibial component changed over 4 degrees by the time of the follow-up, but the clinical knee scores of 6 of the knees were greater than 75 points. The medial angle of the tibial component tended to be 90 degrees at the time of the follow-up study, although the angles had a variation of about 4 degrees.