One of the most difficult problems in orthopaedic oncology is reconstruction after resection of a tumor of the proximal end of the femur. In order to achieve a wide margin about a primary or secondary malignant neoplasm of bone, it is often necessary to resect not only the hip joint and fifteen or more centimeters of the proximal part of the femur, but also the surrounding envelope of soft tissue. In some patients, little is left but the sciatic and femoral nerves and vessels. Since 1971, we have done reconstructions in forty-four patients, using an allograft and an implant or else an osteoarticular graft alone to replace the proximal end of the femur. Twenty-eight of these reconstructions were performed in patients who had had a malignant tumor and were followed for two to fifteen years postoperatively. Fifteen of the patients had only an osteoarticular graft, and thirteen had an allograft and a prosthesis (nine Austin Moore, two bipolar, and two long-stem total hip replacements). The average length of the femoral segment was 18.4 centimeters; the longest one measured thirty-one centimeters. Using an evaluation system of functional end-results that includes failures as a result of recurrence of the tumor, we recorded approximately 70 per cent excellent and good results for both groups. When the two failures that were due to recurrence of the tumor were omitted from the statistics (in order to evaluate the allograft procedure more fully), the successful results increased to about 80 per cent. In general, the patients who had an osteoarticular reconstruction fared less well than did those who had an allograft and a prosthesis, but the series were not quite comparable. The major complications were metastases in nine patients (five of whom died), infection in five, and fracture of the allograft in six. Restoration of the reconstruction was possible for most of the patients who had a problem that was not related to the tumor, and only one patient required an amputation for recurrent tumor. Despite the many difficulties, we think that an allograft, with or without a prosthetic implant, should be given primary consideration as a means of reconstruction of the limb when resection of a tumor necessitates resection of a long segment of the proximal end of the femur.