A clinical study of scoliosis was carried out on 296 cases of structural scoliosis in the department of orthopaedic surgery, Hanyang University Hospital for 26 years and 6 months between Jan. 1963 and June 1989. The results obtained were as follows: l. Idiopathic scoliosis was the most common cause of strctural scoliosis. 2. The ratio between male and female was 1:1.7(110:186). 3. The most common site of scoliosis was the lumbar region for paralytic scoliosis and thoracic region for idiopathic scoliosis. 4. In idiopathic scoliosis, the thoracic curve was convex to the right in 79.1%. The thoracic curve in the combined type with double curvature was convex to the right in the most 80.4% of both idiopathic and paralytic scoliosis. 5. Severe curvatures over 80° was observed in 30.8% of paralytic scoliosis and in 5.6% of idiopathic scoliosis, being more frequent in paralytic than in idiopathic scoliosis. 6. The number of vertebrae involved in the major curve of scoliotic region was between 6 and 9 in most cases of both paralytic and idiopathic scoliosis. 7. Rotation grade was increased with the degree of eurvature in either paralytic or idiopathic scoliosis. Wedging was increased with the degree of curvature in idiopathic scoliosis, but no significant correlation with the degree of curvature was observed in paralytic scoliosis. 8. Of 37 cases of congenital scoliosis, 19 cases(51.4%) of them were caused by hemivertebra and the majorty of this congenital anomaly was located at the lumbar region. 9. The therapeutic effect of correction with Milwaukee brace was better on idiopathic scoliosis (15.8%) than in paralytic scoliosis(9.8%).10. In paralytic scoliosis treated with posterior spinal fusion without instrumentation, the postoperative correction was achieved in 48.7% of the cases studied with the final correction angle of 49.9°and the correction loss was observed in 12.7%. The postoperative correction with combined anterior or posterior spinal fusions with instrumentation was achieved in 51.7% with final correction angle of 53.0° and the correction loss was observed in 7.9%.11. With instrumentation, the postoperative correction rate of the structural curvature was achieved in 49.6% of the cases studied and the correction loss was observed in 5.1%. The best results were obtained in the thoracic curvature with 3.9% of the correction loss. 12. Three cases of postoperative complications was observed in this study: pseudoarthrosis and bending of graft with some loss of correction following the posterior fusions without instrumentation, and displacement of distraction hook on rod with Harrington instrumentation.