The percentage of aseptic loosening of different types of total hip prostheses in 771 cases has been evaluated. The time of follow-up ranged between 1/2 and 6 years. In prostheses of the Charnley-Müller type of loosening was 7.8% in the Charnley-Müller-Weller type with a 2-year shorter follow-up it was 2.4%. No loosening was found during the last 2 1/2 years when the prosthesis of Beck was used, which has an increased femoral shaft-neck angle and a longer stem. When a longer neck was used, the loosening averaged 5.6%. In cases of normal neck-length it was 3.6%. From the biomechanical standpoint it seems preferable to use prostheses with an increased shaft-neck angle. Comparing the rates of loosening as reported in the literature it appears that prostheses where the femur is displaced medially (by a small femoral head or increased shaft-neck angle) show a very low rate of loosening. Models with metal sockets (McKee-Farrar and Weber-Huggler) seem to have a higher rate of loosening than models with polyethylene sockets.