Two experimental bone-graft substitutes, one composed of a porous hydroxyapatite plus an osteoinductive protein and the other composed of lactomer beads (copolymers of polylactic acid and polyglycolic acid) plus an osteoinductive protein, were evaluated as a means of reconstructing a large cranial defect model created in the rabbit. Twenty-five rabbits underwent a 16 x 20 x 1.5 mm full-thickness (extradural) excision of the parietal bone, were divided into different groups of five rabbits each, and were reconstructed by (1) hydroxyapatite, (2) hydroxyapatite plus protein, (3) lactomer beads, and (4) lactomer beads plus protein, and (5) one group consisted of nonreconstructed controls. The implants were harvested at 12 weeks and analyzed for percentage of bone ingrowth by histologic examination of decalcified mid-coronal sections stained with hematoxylin and eosin. The groups reconstructed with protein-treated implants demonstrated significantly greater amounts of ingrowth than those with untreated implants. Both protein-treated hydroxyapatite and lactomer bead groups had more than twice the amount of bone ingrowth than their respective untreated groups (29.0 versus 12.8 percent, p < 0.005, and 27.0 versus 10.0 percent, p < 0.001, respectively). The new bone found in the protein-treated and untreated implants was quite different: lamellar and woven, respectively. The results strongly suggest a clinical role for the combination of the mechanisms of osteoconduction and osteoinduction in the treatment of bone defects.