There have been many debates concerning operative decompression of treatment of thoracolumbar burst fractures with retropulsed bone fragment. From March 1988 to February 1992, authors treated thirty-three thoracolumbar burst fractures by using transpedicular screw fixation and posterior fusion via the posterior approach. We attempted to reduce retropulsed fragment by ligamentotaxis alone and not to do posterolateral nor anterior decompression. After the reduction of fractured spine by posterior instrumentation, we tried to determine the efficiency of reduction of the retropulsed fragment by ligamentaxis along. As a method, we compared the change of anteroposterior, transverse to diameter and area of spinal canal of fractured spine between preoperative and the postoperative situation. The results were as follows; 1. The mean anteroposterior and transverse diameter of the spinal canal on computed tomogram film was 10.1mm & 21.8mm preoperatively & 12.4mm & 23.2mm postoperatively, showing an increase. 2. The area of spinal canal of involved spine on CT film was evaluated preoperatively & post-operatively, the mean spinal canal invasion rate decreased from 36.3% preoperatively to 14.3% postoperatively. 3. The degree of reduction of middle height on plain x-ray and reduction of spinal canal invasion on computed tomogram were statistically correlated(p<0.01). 4. There was no correlation between the degree of canal narrowing and degree of neurologic impairment. also, there was no correlation between the reduction of retropulsed fragments and subsequent neurologic impairment. 5. There was the relatively satisfactory enlargement of the spinal canal on computed tomogram at the follow-up So we suggest that it is possible to get enough decompression through reduction of retropulsed fragment by ligamen to taxis alone without posterolateral decompression.