This study was designed to determine which of several bone grafting materials would be the most efficacious substitute for autogenous bone graft in the treatment of segmental long bone defects. The experimental model was a 1-cm defect in the rabbit ulna. The control group had nothing implanted in the defect. The six grafts tested were: (a) autogenous iliac crest bone, (b) autogenous cortical bone (ulna), (c) hydroxylapatite, (d) hydroxylapatite-demineralized bone matrix (allograft) composite graft, (e) freeze-dried bone (allograft), and (f) demineralized bone matrix (allograft). At 6 weeks postoperatively, the ulnas were harvested, examined radiographically, and tested mechanically in torsion. The radiographic examination proved to be of little value because some materials were radiodense at the time of implantation. The rates (percentage) of union, torques at failure, and energy to failure values were statistically significantly higher than control for all groups except hydroxylapatite. We concluded that demineralized bone matrix and hydroxylapatite-demineralized bone matrix composite graft compare favorably with cortical replacement (autograft) in mechanical strength and rate of union and therefore may be satisfactory substitutes for bone grafting. Freeze-dried bone did not appear to be as satisfactory because of its low mean energy to failure, but statistical analysis failed to confirm this opinion. Hydroxylapatite graft, when used alone, does not appear to be a suitable material for grafting segmental bone defects.