Of fifty-seven adult dogs in which standard bilateral mid-ulnar fractures were produced, thirty-four had one side fixed with either a tight-fitting intramedullary Steinmann pin or a four-hole plate while the other side was not fixed, and twenty-three had a plate on one side and an intramedullary rod on the other. The blood flow at the fracture site, as determined by 85Sr clearance, was significantly less on the rod-fixed side than on the plate-fixed side at fourteen and ninety days, while the total ulnar blood flow was significantly higher on the rod-fixed than on the platefixed side at one and fourteen days but was the same on both sides thereafter. When the plate-fixed and rod-fixed fracture sites in the same dog were compared with respect to periosteal and endosteal bone formation as determined by tetracycline labeling anc microradiography, bone formation, like the blood flow, was greater in the callus of the ulnae fixed by plates. However, at ninety days, when healing had advanced to a point when the extent of union could be judged clinically, union had occurred in both ulnae of six of the seven dogs with both types of fixation studied at this time and was delayed in both ulnae of the remaining dog.