The effects of early physiologic dynamic compression on fracture healing were studied in the dog. Transverse midtibial osteotomies were performed bilaterally and stabilized with a relatively rigid external fixation system in a neutralization mode (800 microns) to prevent compression of the osteotomy ends during weight bearing. On the 15th day, one osteotomy in each animal was subjected to dynamic compression through weight bearing by release of the fixator-telescoping mechanism (axial dynamization), while the other side remained unchanged as the control. Analysis of sequential roentgenograms showed that the callus distribution was more symmetric on the dynamic compression side. The two sides showed no significant differences in quantitative technetium-99 bone scans or in osteotomy-site blood flow. There were no statistical differences in new bone formation, bone porosity, or maximum torque between sides. The fixation had maintained the initially created osteotomy gap on the control side and tended to unite through a gap-healing mechanism. The dynamic compression side showed reduction in gap size and union by more of a contact-healing mechanism. There were no statistical differences in the rate of pin loosening, but its distribution according to pin location was significantly different between the two sides.