To assess the effect of sodium fluoride (NaF) in the healing of a defect in cortical bone, an experimental model was created by the drilling of 5.0 mm holes in the proximal ends of both femora of 12 adult male New Zealand White rabbits. An interlocking intramedullary implant constructed of poly(d,l-lactic acid) containing NaF was placed in the right femur and an identical implant without NaF (sham), in the left. The implant in the right femur was designed to release NaF in a controlled manner over the duration of the experiment. Ten weeks after implantation, the specimens were removed and were tested in torsion. The mechanical properties were not significantly different between the groups. The femora exposed to NaF had an 18.6% increase in intact cortex near the defect (p = 0.023), however, the deposition of mineralized bone within the defect was not significantly greater. In fact, healing appeared to be impaired by the presence of NaF. There was complete closure of the defect in all but one of the femora with a sham implant, but the tissue had not yet calcified. In contrast, only one femoral defect exposed to NaF had closed. Examination of the material filling the defects of the femora exposed to NaF showed that it was predominantly uncalcified osteogenic mesenchymal tissue.