We used two mechanically dissimilar devices, compression bone-plates and unilateral-frame external fixators, in a standard canine osteotomy model; both methods were highly successful in achieving mature bone union. Bone union was studied by histological, physiological, and biomechanical means. At 120 days after injury, union was biomechanically less mature on the external fixator side. These tibiae had less intracortical new-bone formation (p less than 0.01), more bone resorption (p less than 0.045), and more bone porosity (p less than 0.04) when compared with paired tibiae that had been treated with compression plates. This higher level of bone turnover on the external fixator side was accompanied by an increase in blood flow (measured by clearance of 85Sr) (p less than 0.04). At the osteotomy site, pre-experimental or unlabeled bone and porosity were greater on the external fixator side and endosteal new-bone formation was greater on the plated side. Since the in vitro stiffness of the external fixator was less in all modes tested (compression, distraction, torsion, and anteroposterior bending) except lateral bending, it may be that the rigidity of the fixation is an important factor in early bone-remodeling of a healing osteotomy.