We report a clinical and radiological survival analysis of 148 consecutive revisions of hip arthroplasties. All patients referred were offered revision if it was indicated and they were medically fit. About one-third (32%) had had at least one previous revision, and about one-third had an established deep infection. The revisions were performed in ultra-clean air with body-exhaust suits. The usual method of fixation was by contained and pressurised cement of standard viscosity, to which appropriate antibiotics had been added in infected cases. Some patients had cementless revision. Clinical failure meant that one or both of the implants had been removed; radiological failure was assessed from serial radiographs. The clinical survival at ten years was 95%. The Merle D'Aubigne and Postel rating for pain improved from a mean of 2.9 to 5.2; and in Charnley group A and B cases, walking ability improved from 2.3 to 4.3. In contrast to some reports we also found good radiological survival; this was 90.5% for cemented femoral stems. Isoelastic revision stems inserted without cement gave poor results.