One hundred and thirty-eight consecutive revisions of an acetabular component were performed in 132 patients between 1983 and 1986. The revision prosthesis was a hemispherical component that was coated with porous titanium mesh and was secured to the pelvis with a variable number of screws. There were seventy-five women and fifty-seven men, and the mean age at the time of the revision was fifty-two years (range, twenty to seventy-nine years). Due to defects in the acetabulum, 80 per cent of the hips were treated with bone grafts, usually a mixture of local autogenous graft and freeze-dried allograft. One hundred and twenty-four patients (129 hips) were available for follow-up, at a mean of forty-four months (range, thirty-six to eighty months). Seven hips (5 per cent) were revised again: four because of infection and three because of instability. There were no revisions for loosening without infection, and none of the components migrated in the absence of infection. Radiolucent lines were common and usually corresponded to regions in which allograft had been used. All bone grafts united by twelve months. Non-contained medial grafts underwent a phenomenon of resorption, consolidation, and remodeling to a sclerotic rim by twenty-four months. There were no complications related to placement of the screws. The results in this series were superior to results of revisions of acetabular components with the use of cement, which were associated with rates of failure of almost 10 per cent after a similar duration of follow-up. The technique of maximization of host-bone coverage and of packing of all defects with cancellous autogenous graft or allograft, or both, was successful for all classes of acetabular deficiencies.