Sixty-nine patients (seventy-eight elbows) who had rheumatoid arthritis were managed with a Coonrad-Morrey total elbow arthroplasty between 1981 and 1986. At the time of the present review, forty-one patients (forty-six elbows) were alive and had been followed for at least ten years after the procedure (Group 1). The remaining twenty-eight patients (thirty-two elbows) had died or had had a revision less than ten years after the procedure or had been followed for less than ten years (Group 2). The patients in Group 1 had a younger mean age at the time of the procedure, but all other preoperative parameters were similar for both groups. At the latest follow-up evaluation, 97 per cent of the elbows (forty-five of the forty-six in Group 1 and thirty-one of the thirty-two in Group 2) were not painful or were only mildly painful. The mean arc of flexion-extension was 28 to 131 degrees; this represents an increase of 13 degrees (15 degrees in Group 1 and 7 degrees in Group 2) compared with the preoperative value. The mean arc of pronation was 68 degrees, and the mean arc of supination was 62 degrees; this represents an increase of 21 degrees. The results for seventy-four of the seventy-eight elbows (all forty-six in Group 1 and twenty-eight of the thirty-two in Group 2) were considered satisfactory by the patients. One patient thought that the status of the elbow was unchanged compared with preoperatively, and three thought that it was worse. Seventy-six of the seventy-eight elbows had long-term radiographic evaluation; the two remaining elbows were excluded because a resection arthroplasty had been performed. There were two loose ulnar components; one was associated with an infection, and the other had been causing no symptoms at the time of the patient's death. In addition, both components were radiographically loose in an elbow that had had a revision without cement after a previous total elbow arthroplasty. Five bushings (7 per cent) were completely worn, and six (8 per cent) were partially worn. Complications occurred in eleven elbows (14 per cent) and were serious, necessitating reoperation, in ten (13 per cent). Delayed complications included three avulsions of the triceps, two deep infections, two ulnar fractures, and one fracture of an ulnar component. In addition, two elbows were revised because of aseptic loosening. No patient had persistent ulnar neuritis or serious skin complications. At the latest clinical follow-up evaluation, according to the Mayo elbow performance score, forty-three of the seventy-eight elbows had an excellent result; twenty-six, a good result; seven, a fair result; and two (both in Group 2), a poor result. The rate of survival of the prosthesis was 92.4 per cent, with 86 per cent good or excellent and 14 per cent fair or poor results.