Objectives: This study was designed to evaluate the effect of gap width and graft placement on bone healing around implants placed in simulated extraction sockets of various widths in four Labrador dogs. Materials and Methods: Five Osseotite (R) implants per dog were placed in the mandible of four dogs. Two implants were inserted into sites with a 2.37 mm and two with a 1 mm gap present between the implants and bone around the coronal 6 mm of the implants in each dog. For one of each gap sizes, the gap was filled with Bio-Oss (R), and the other two with blood alone. A fifth implant was inserted without a gap and used as a control. Ground sections were prepared from biopsies taken at 4 months and histometric measurements of osseointegration and bone between the threads made for the coronal 6 mm. Results: The medians for osseointegration ranged from 5.2 mm for control to 1-2.6 mm for the test modalities. There were significant differences for linear measurements of osseointegration ((2)(chi) 18.27; df 4; P = 0.0011) and bone area within threads (chi(2) 23.4; df 4; P = 0.0001) between test modalities. Conclusions: The results suggest that the wider the gap around the implants, the less favourable the histological outcome at short time intervals following treatment. They also infer that bone grafting with anorganic bovine bone xenograft seems to lead to a more favourable histological outcome for wider circumferential defects but not for narrower defects.