The value of histocompatibility matching in frozen bone allografts was studied in a canine cancellous ulnar segmental-replacement model. Frozen bone that was exchanged across strong and weak transplantation barriers was evaluated histologically and radiographically at thirteen and twenty-six weeks after grafting. Histological grading criteria quantified the type of union at each end of the graft and the degree of remodeling of the marrow, spongiosa, and compacta. Radiographic grading criteria included the presence of union at each end of the graft and the degree of remodeling of the graft segment. In vitro studies for serum antibody and cell-mediated immunity were carried out by isotopic cytotoxicity methods at seven intervals during the twenty-six-week study period. Histologically, the strong-barrier allografts had fewer osseous unions and less reorganization of spongiosa and marrow when compared with autograft controls at both thirteen and twenty-six weeks. Radiographically, the strong-barrier allografts at thirteen weeks had fewer unions and marked resorption of grafts material when compared with autograft controls. There were no differences between weak transplantation-barrier grafts and control autografts radiographically or histologically at thirteen and twenty-six weeks after grafting. Frozen bone allografts did not elicit detectable serum antibody or lymphocytes that were cytotoxic for donor cells.