Purpose: The implant-abutment connection (microgap) influences the peri-implant bone morphology. However, it is unclear if different microgap configurations additionally modify bone reactions. This preliminary study aimed to radiographically monitor peri-implant bone levels in two different microgap configurations during 3 months of nonsubmerged healing. Materials and Methods: Six dogs received two implants with internal Morse taper connection (INT group) on one side of the mandible and two implants with external-hex connection (EXT group) on the other side. One implant on each side was positioned at bone level (equicrestal); the second implant was inserted 1.5 mm below the bone crest (subcrestal). Healing abutments were attached directly after implant insertion, and the implants were maintained for 3 months without prosthetic loading. At implant placement and 1, 2, and 3 months, standardized radiographs were taken to monitor peri-implant bone levels. Results: All implants osseointegrated. A total bone loss of 0.48 ± 0.66 mm was measured in the equicrestal INT group, 0.69 ± 0.43 mm in the equicrestal EXT group, 0.79 ± 0.93 mm in the subcrestal INT group, and 1.56 ± 0.53 mm in the subcrestal EXT group (P > .05, paired t tests). Within the four groups, bone loss over time became significantly greater in the EXT groups than in the INT groups. The greatest bone loss was noted in the subcrestal EXT group. Conclusion: Within the limits of this animal study, it seems that even without prosthetic loading, different microgap configurations exhibit different patterns of bone loss during nonsubmerged healing. Subcrestal positioning of an external butt joint microgap may lead to faster radiographic bone loss.