Because no report has been issued on the healing effects of low-intensity pulsed ultrasound on moderate to large fracture gaps, we performed an experimental study using acanineulna full-defect model. Ten mongrel male dogs were divided into two groups: a small defect group and large defect group. The defects were made on the middle one third of both ulnae and one side only was randomly selected for ultrasound sonication, at 1 MHz, 200 microsecond bursting sine wave in 50 mW/cm² spacial average and temporal average. Sonication was started on the day after surgery and applied for 15 minutes once a day for six days a week. In the small defect group, the means of the radiologic scores, as described by Lane and Sandhu, were 0.6, 4.4, and 8.4 in the control side and 1.8, 6.0, and 10.4 in the treatment side one, three, and five months after the operation, respectively (p=0.0372). In the large defect model, the corresponding means were 2.2, 3.4, and 6.0 in the control side and 3.3, 5.4, and 9.2 in the treatment side (p=0.009). Low-intensity pulsed ultrasound enhanced new bone formation in small and large full-defects and decreased the incidence of nonunion in the large defect model.