The loss of multiple teeth in the posterior mandible often results in deficient ridge width for prosthetic tooth rehabilitation. Multiple approaches have been used to regenerate lost bone; however, the outcomes are highly variable. Several approaches use combination therapy to make the procedure more predictable. In this experimental study in dogs, a chronic defect was created and then treated with one of four therapies: (1) autograft, (2) recombinant human bone morphogenetic protein 2 (rhBMP-2) in absorbable collagen sponge (ACS), (3) ACS wrapped around a collagen-ceramic matrix, and (4) rhBMP-2 in ACS around the collagenceramic matrix. Two metal screws were used for space maintenance. After 2 months, the metal pins were removed and an endosseous dental implant was placed in the regenerated bone. Two months later, the animals were sacrificed and specimens prepared for histologic evaluation. Only five animals were used, with each condition evaluated in each animal. With this low number of animals and with the observed variability, no quantitative differences were found between each of the four conditions evaluated. All conditions resulted in new bone growth. Dense native bone was found in the autograft sites and the sites treated with rhBMP-2 and ACS. Sites treated with the collagen-ceramic matrix with and without rhBMP-2 in the ACS had residual ceramic and large porous areas. Bone was found in varying degrees along the implant surfaces. These results suggest that multiple approaches can be used to augment bone horizontally in the posterior mandible of dogs. Interestingly, rhBMP-2 combined with a non-space maintaining collagen carrier yielded new bone densities similar to the autograft in this model.