One hundred and forty revisions for loosening of the acetabular component were performed in 124 patients, with the use of an uncemented Harris-Galante acetabular prosthesis. A component was considered loose if there was a change in its position or vertical or horizontal migration, or both, of four millimeters or more, as demonstrated on serial radiographs. All patients were followed prospectively for a mean of forty-one months (range, twenty-four to seventy-six months). Bone-grafting was performed at the time of the revision in 127 of the hips. Identifiable failure of fixation of the acetabular component occurred in only two hips (1 per cent), which had both severe acetabular bone loss and pelvic discontinuity. In one of these hips, fixation of the component could not be achieved during the revision, and the component subsequently migrated. No other components migrated. Only one patient had a revision of the index acetabular operation. A continuous radiolucency developed at the bone-mesh interface of five acetabular components, and in one other hip a small portion of the mesh separated. This hip was reoperated on for a problem with the femur, and the socket was found to be rigidly fixed. No other evidence of loosening was identified. All of the bone grafts united, but partial resorption of the graft occurred in thirty-nine hips. Although there were substantial osseous defects that necessitated major bone-grafting, revision with the uncemented Harris-Galante porous-coated acetabular component provided superior fixation compared with that reported in other series in which cemented acetabular components were used for revision. Of the 140 hips, eighty-nine (64 per cent) had a postoperative score of good or excellent, according to the Harris hip-rating system. Twenty-eight (70 per cent) of the forty hips that had revision of the acetabular component alone were rated as good or excellent.