One hundred and six consecutive total hip arthroplasties with cement were performed by one surgeon, at least ten years before the time of the present clinical and radiographic review, in seventy-five patients who had adult-onset rheumatoid arthritis. Two patients (three hips) were lost to follow-up. Seven (7 per cent) of the remaining 103 hips were revised. The revisions were performed because of infection (three hips), dislocation (two hips), or aseptic loosening (two hips). Of the ninety-eight hips that were not lost to follow-up or revised because of infection or dislocation, eight (8 per cent) had radiographic loosening of the acetabular component and two (2 per cent) had radiographic loosening of the femoral component. Although the prevalence of radiographic loosening of the acetabular component was four times greater than the prevalence of radiographic loosening of the femoral component, the prevalence of revision because of aseptic loosening of the acetabular component was identical to that for the femoral component (one component each). These results compared favorably with those of total hip arthroplasty with cement, performed by the same surgeon, for the treatment of other diagnoses. Loosening of the acetabular component was significantly associated with a younger age at the time of the index operation (p = 0.03) and with acetabular osteolysis (p = 0.0006). Of forty-eight hips in thirty-two patients who survived for at least ten years, 96 per cent (forty-six hips) were considered by the patients to have a satisfactory result. At the time of the latest follow-up, twenty-four (75 per cent) of the patients had no pain in the hip. Although eighteen patients (56 per cent) could walk without support at a minimum of ten years after the operation, we found that the functional results for patients who had rheumatoid arthritis were inferior to those observed for patients who had had a total hip arthroplasty with cement, performed by the same surgeon, for the treatment of other diagnoses.