Various methods and graft materials have been used to fill in the defect adjacent to the implants and considered as clinically acceptable. But it is not clear whether the regenerated bone increases the implant-bone contact and supports the implant. The purpose of this study is to evaluate regenerated bone surrounding implants using bone morphogenetic protein(BMP) and demineralized freeze-dried bone(DFDB), and the interfaces between implants and regenerated bone. bBMP was extracted and partially purified from the bovine bone matrix using heparine chromatography. Demineralized freeze-dried bone was made from the dog. Inactive insoluble collagenous bone matrix(IBM) of dog was used as carrier of bBMP. Interfaces of titanium coated epoxy resin implants were processed for demineralized section for transmission electron microscopy(TEM) and those of screw type implants were for nondemineralized section for light and fluoromicroscopic examination. Implants were inserted in the inferior border of mandible of adult dogs and artificial bony defects (3x3x4mm) were made at the mesial and distal side of implants. Defects were filled with BMP(BMP group) and DFDB(DFDB group). For the fluoromicroscopic examination, the fluorescent dyes (oxytetracycline, calcein green, alizarin red) were injected 2, 4, 6, 8, 12 weeks after implantation. The experimental animals were sacrificed at the 6th and the 12th week and their mandible were extirpated and processed for examination with light microscopy, fluoromicroscopy and TEM. The obtained results were as follows : 1. By the light microscopic findings, the defects were filled with woven bone at the 6th week and compact bone at the 12th week, and the osseointegrations were seen in both groups. There was no histological difference between them. 2. On the basis of the histomorphometric analysis, BMP group (6th week: 40.25%, 12th week: 56.04%) had higher bony contact ratio than DFDB group (38.37%, 42.63%). There was significant difference between two groups at the 12th week (P<0.05). 3. The amount of bone formation in BMP group was more prominent than in DFDB group. Significant difference was noted among two groups at the 6th and the 8th week (P<0.05). 4. By the transmission electron microscopic findings, 0.4-2micrometer soft tissue layer was found in adjacent to the interfaces and over the collagen fibrils of bone at the 6th week. However, about 100nm amorphous layer was noted at the interface or collagen fibrils directly extended to the titanium surface at the 12th week. There was no significant difference between two groups. 5. These results suggest that BMP and DFDB can be used as good graft materials in the regeneration of bone adjacent to implant, and BMP is more valuable as a bone inducer than DFDB.