One hundred and fourteen total hip arthroplasties in 100 patients were performed using the CAD (computer-assisted design) prosthesis at The Hospital for Special Surgery between 1975 and 1977. Sixty-one of these patients (seventy-four hips) have been followed both clinically and radiographically for a minimum of five years (average, 7.1 years). While no hip required revision before five years, seven hips were revised between six and ten years postoperatively. Ten hips had an excellent result; forty-eight, good; eight, fair; and one, poor. Radiographic analysis revealed narrowing of the cortex of the bone medially at the region of the proximal part of the stem in the hips that were operated on as compared with the contralateral side that was not operated on. Nine hips (12 per cent) had one to two-millimeter radiolucent lines at the femoral bone-cement interface, and in five the lines were progressive. Fourteen hips (19 per cent) had radiolucent lines at the femoral bone-cement interface with a maximum width of 0.5 millimeter. Thirty-four (42 per cent) had more than three millimeters of resorption of the calcar or superomedial cyst formation. Survivorship analysis of the prosthesis predicted a 77 per cent survival rate at nine years. None of the prosthetic stems fractured despite the fact that the population was young, active, heavy, and predominantly male. The incidence of calcar resorption, however, was higher than in other studies. Thus, while the revision rate compared favorably with that of similar studies using other stem designs, compromise of the bone stock of the proximal end of the femur may be a significant disadvantage of this type of prosthesis.