One hundred and thirty-one patients who had 144 cemented or uncemented hip prostheses were followed prospectively for two to four years. A cemented or a hybrid prosthesis (consisting of a cemented femoral component and an uncemented acetabular component) was used in men older than seventy years, in women older than sixty years, and in younger patients in whom adequate initial fixation could not be obtained without cement. Uncemented, porous-surfaced implants were used in all other patients. The over-all clinical results were similar for the three groups. For the fifty-two hips that had a cemented prosthesis, the mean total Harris hip rating was 91 points and the score for pain, 42 points; for the twenty-seven hips that had a hybrid prosthesis, 90 and 43 points; and for the sixty-five hips that had an implant allowing ingrowth of bone in both the acetabulum and the femur, 95 and 43 points. Two prosthetic stems that were designed to allow ingrowth of bone had aseptic loosening; one was revised. Pain in the thigh, usually slight and not disabling, occurred at one year in 24 per cent of the patients in whom a femoral component allowing ingrowth had been used; the prevalence of pain then declined. The incidences of migration of the components and of radiolucent lines were greater in the acetabula that had a cemented component than in those that had a cup allowing ingrowth of bone.