Over the past 24 years, the authors have implanted >870 massive frozen cadaveric allografts mostly for the treatment of defects created by the resection of a bone tumor. Most of the grafts were obtained from the authors' institutional bone bank. The results show that only stage and type of graft affected outcome predictably. Specifically, grafts for a Stage 2 or Stage 3 tumor had a poorer outcome than those for Stages 0 and 1. The results for allograft arthrodeses were considerably poorer than osteoarticular, intercalary, and allograft plus prosthesis. The other major factors in results were complications--recurrence, infection, fracture, and nonunion--with the former 2 having a profound negative effect on outcome. After the first year of susceptibility to infection (10%) and the third year of increased risk of fracture (19%), the grafts become stable, and approximately 75% are retained by patients and are considered to be successful for >20 years after implantation. Osteoarthritis becomes a problem at approximately 6 years for osteoarticular grafts, and so far, 16% of the patients with distal femoral, proximal tibial, or proximal femoral grafts have required total joint replacements. Although the current results are adequate, they are imperfect, and research should be directed at improving the results.