The results of 140 total hip revision procedures for "non-septic" loosening, dislocation, and fracture of the femoral stem or shaft have been personally reviewed and rated by the Harris method. The minimum follow-up period was six months: thirty-three (24 per cent) showed excellent or good results, seventy-two (51 per cent) showed fair or poor results. Subsequent excision arthroplasty was performed in thirty-one patients. The infection rate for these revision procedures was very high, suggesting that many were already infected at the time of revision, and that every "loose" hip must be assumed to be infected until proved otherwise. The mortality rate of 3 per cent was surprisingly low after more than one major surgical procedure in these elderly patients.