Two hundred and ten hips in 206 patients who had an initial total hip arthroplasty performed at the Mayo Clinic between 1969 and 1978 required revision of the arthroplasty at the Mayo Clinic for reasons other than infection. One hundred and sixty-two of the patients (166 hips) were followed both clinically and roentgenographically for two years or more. One hundred and forty-five (90 per cent) reported that they had improvement after the surgical revision. Complications that occurred with revision included deep sepsis, superficial would infection, dislocation, intraoperative femoral fracture, and postoperative femoral fracture. Roentgenographic analysis showed probable loosening in thirty-three acetabular components (20.1 per cent) and seventy-two femoral components (44 per cent). Symptomatic loosening (moderate to severe pain and probable roentgenographic loosening) was seen in thirty-five patients. Eight patients required a second revision for this reason, and seven others required a second revision for other reasons. Modified Harris hip scores, calculated for 108 hips, showed a good or excellent result in sixty-seven hips (62 per cent), a fair result in twelve (11 per cent), and a poor result in twenty-nine (27 per cent). Using a new Mayo Clinic hip score that incorporates roentgenographic data (which will be described) in the evaluation of 165 revised hips, there was a good or excellent result in eighty-five (52 per cent), a fair result in thirty-two (19 per cent), and a poor result in forty-eight hips (29 per cent). Although 90 per cent of the patients thought that their condition had improved, the high incidence of roentgenographic signs of probable loosening of a component is of serious concern.