Abstract OBJECTIVES: To report an ovine model that can be used to evaluate the efficacy of bone substitutes for repair of segmental diaphyseal bone defects. STUDY DESIGN: Experimental study. ANIMALS: Eleven 2-year-old Pré-Alpes Sheep. METHODS: Mid-diaphyseal metatarsal bone defects (25 mm long) were stabilized by a dynamic compression plate over a polymethylmethacrylate (PMMA) cement spacer, and by external coaptation. The PMMA spacer was removed at 6 weeks by incising the encapsulating membrane. The defect remained unfilled (Group 1; n=5) or was filled with morselized autologous corticocancellous graft (Group 2; n=6), the membrane sutured closed, and external coaptation applied for 6 months, when healing was evaluated. RESULTS: Radiographic, computed tomographic, and histologic examinations at 6 months after the 2nd surgery revealed non-union in ungrafted defects whereas grafted defects showed bone healing. The induced membrane had blood vessels, CBFA1+ cells, and very few macrophages entrapped in a collagenous tissue positive for type I collagen. CONCLUSION: This ovine metatarsal defect model resulted in a critical-size defect (non-union) that healed when grafted. The PMMA-induced membrane constrained the graft, was well vascularized, and may have osteogenic properties. CLINICAL RELEVANCE: This model may be useful to evaluate new strategies in bone tissue engineering because the PMMA-induced membrane may help confine bone morphogenetic proteins, skeletal stem cells, or other agents to the defect cavity where they could be useful to enhance bone formation.