The purpose of this study was to evaluate the effect of gap width on bone healing around implants placed into simulated extraction socket defects of varying widths in 10 mongrel dogs. All premolars were removed and the alveolar ridges were reduced to a width of 7 mm. Nine weeks later, a total of 80 implants, 10 mm long by 3.3 mm wide, were placed into osteotomy sites prepared to 3 different diameters in the coronal half, simulating extraction sockets. Three experimental sites, with gap sizes of 0.5 mm, 1.0 mm, and 1.4 mm, were created; the control sites had no gap. The depth of each defect was measured at the time of implant placement. All implants were stable at the time of placement. The dogs were euthanized 12 weeks after implant placement, and blocks containing the implants and adjacent bone were submitted for histologic evaluation. Clinically, all control and test sites healed, with complete bone fill in the defect. Percentages of bone-to-implant contact were measured histologically. As the gap widened, the amount of bone-to-implant contact decreased, and the point of the highest bone-to-implant contact shifted apically. These changes were statistically significant (P < .001). No statistically significant differences in bone-to-implant contact were found between the sites when the apical 4 mm of implants were compared. Within the limits of this study, the simulated extraction socket defects healed clinically, with complete bone fill, regardless of the initial gap size. However, the width of the gap at the time of implant placement had a significant impact on the histologic percentage and the height of bone-to-implant contact.