We compared the radiographic results of secondary total hip replacements, 99 following failed uncemented hemiarthroplasties and 21 following failed mould arthroplasties, with those of 825 primary cemented total hip replacements. The probability of occurrence of a number of radiological changes over time was calculated using survival analysis. The mean follow-up was 7.6 years (range one month to 20 years). The performance of the secondary total hip replacements varied with the preceding implant and was different for acetabular and femoral components. The incidence of radiological loosening was higher for femoral components implanted after failed hemiarthroplasties and for acetabular components after failed mould arthroplasties. However, the incidence of continuous radiolucent lines was lower for the acetabular components of converted hemiarthroplasties than for the primary replacements.