This article reports a comparison of the Cotrel-Dubousset (CD) operation with the Harrington-Luque (HL) procedure for the treatment of adolescent idiopathic thoracic scoliosis. Thirty-nine patients were studied preoperatively and postoperatively using segmental radiologic measurements including Cobb angle, end-vertebra angles (EVAs), surgical flexibility index, vertebral rotation, displacement and tilt, convex and concave rib-vertebra angles (RVAs), and kyphosis and lordosis. Cotrel-Dubousset is not significantly different from HL with respect to Cobb angle (%), surgical flexibility index, apical vertebral displacement, apical vertebral rotation, apical rib-vertebra angles, kyphosis, and lordosis. It is significantly better than the HL with respect to the correction of vertebral displacement at T10-11; lower EVA of the thoracic curve; vertebral rotation mainly above the apex; convex RVAs above the apex; and concave RVAs at T10. The surgical correction of vertebral tilt above and below the apex of the thoracic curve is significantly related to the correction of convex and concave RVAs. The view that persistent deformity of ribs is a factor needing surgical correction in some patients with adolescent idiopathic thoracic scoliosis having posterior instrumentation and fusion needs further evaluation.