A stable cortical bone fracture model was developed to evaluate the remodeling rate of cortical bone grafts. Samples of cortical bone were harvested with a trephine and press fit into predrilled holes in the femoral diaphyses of four live dogs. The percentages of new bone, unremodeled graft bone, porosity, forming bone surface area, and resorbing bone surface area were determined morphometrically and compared in cortical autografts, cortical allografts sterilized with 84% ethylene oxide (EO), and allografts sterilized with 12% EO. The host-graft interfaces healed without formation of fibrous tissue or cartilage, indicating a stable fracture surface. The amount of new bone formed in cortical autografts and allografts sterilized with 84% EO was significantly greater than the amount of new bone in allografts sterilized with 12% EO. There was no significant difference between the amounts of new bone formed in the allografts sterilized with 84% EO and the cortical autografts. No significant differences were detected in percentages of porosity or bone surface areas.