Purpose: To study bone healing at implant sites in simulated extraction sockets with 1-mm marginal defects and compare healing around a turned surface (T) to that around a porous oxide surface prepared by anodic oxidation (AO) with or without the use of an autogenous bone graft. Materials and Methods: All mandibular premolars and first molars were extracted from 10 mongrel dogs. After 9 weeks, four sites were prepared on both sides of all mandibles. Each osteotomy was widened in the coronal 5 mm to create a marginal defect of 1 mm around the implants. Autogenous bone was collected during the drilling procedure. The sites were randomized to receive implants with a T or an AO surface, with or without bone grafting. The animals were sacrificed 4 months after implant placement for histologic analysis. Results: Clinically, all sites healed with complete bone fill. The combination of an AO implant and a bone graft resulted in a significantly greater percentage of bone-to-implant contact (BIC) (P .05) versus all other groups. The highest point of BIC was achieved with the AO group, which was significantly greater than the lowest group (T). No significant differences between groups were found when the apical 4 mm (nongap areas) were compared (P = .65). Conclusions: Studies have demonstrated that bone can fill in a marginal defect around a titanium implant with varied histologic BIC, depending on implant surface type and defect dimensions. Based upon this animal study using 10 mongrel dogs, marginal circumferential defects of 1 mm showed significantly higher BIC values for implants that were prepared by AO compared to implants with a turned surface. The addition of autogenous bone grafts further enhanced the degree of BIC. Keywords: anodic oxidation, autogenous, bone graft, histomorphometry, osseointegration