Of ninety-three arthrodeses of the knee joint performed at the Campbell Clinic, sixty-three followed the so-called conventional methods and thirty the compression method. The average time required for clinical and roentgenographic union in the series with the compression technique was less than half as long as that in the series with conventional arthrodesis. Ambulation, too, was possible much earlier in the compression series. In patients with difficult conditions such as tuberculosis, union was obtained by the compression method. With the conventional method, union occured in only 75 per cent and required nearly twice as long. Only four instances of Charcot's disease occurred in the series. In the two patients with compression fixation union was slow, while the ones treated by the conventional methods had non-union. In patients with severe fractures involving the knee joint a longer time for union was required than the over-all average. This delay was rsbab p duhly lme to thie asc Srosis’OSiS which developed as a result of the trauma and impairment of the blood supply. Chapechal's method of intramedullary fixation across the knee joint has not been tried at the Campbell Clinic; however, the principles of his method are in accord with the contemporary concepts of the requirements for arthrodesis and fracture healing. In the one patient treated by the Henderson technique (parallel Knowles pins), the joint united promptly. When the compression apparatus is correctly used in arthrodesis of the knee, the patient is spared much morbidity, discomfort, and disability. It is evident that compresssion arthrodesis of the knee joint possesses many advantages over the conventional method, both to the patient and to the surgeon.