The basic principles for scoliosis surgery learned during the Harrington era are still valid today. Experience has confirmed the need for careful selection of the vertebrae to be instrumented, the value of anterior release for rigid curves in imparting convertibility of the deformity, and the importance of careful fusion techniques. During the last decade, further development has occurred because of an increased knowledge of the biomechanical needs for the internally instrumented spine and a three-dimensional appreciation of the scoliotic curve. Biomechanical advances have centered on an understanding of the load-sharing properties afforded by the multiple spinal purchase sites (segmental spinal instrumentation) and the value of two-rod systems linked by couplers. These advances have provided an increased stiffness of the instrumental spine, a reduction in correction loss, improved fatigue properties of the implant, and fewer pseudarthroses. The most important advance of the last decade is an improved awareness of the three-dimensional approach to the scoliotic deformity with the need to preserve or improve sagittal contours. In particular, the importance of the loss of normal thoracic kyphosis or lumbar lordosis has been emphasized. These conceptual gains have led to the development of many new instrument systems to correct deformity. Each is associated with advantages, problems, and risks that must be understood to make intelligent choices for treatment.