We evaluated the influence of the penetration of wires into the neural canal during segmental wiring in a three-part study. First we examined the anatomy of the thoracic spine, specifically the laminar and interlaminar dimensions, as well as the epidural space. In the second part, we evaluated the depth of penetration of wires into the spinal canal at the time of their passage during spinal segmental instrumentation, using direct laboratory measurements for three configurations of the wire: first with a straight wire, and then with two wires of varying curvature. The measurements were repeated after removal of a portion of the lamina. In the third and final part of the study, we assessed the relationship between the observed penetration of the wires and the depth of penetration as calculated using mathematical models for the three wire configurations. When a wire with the largest possible diameter of curvature was passed under the lamina, there was significantly less penetration using the curved-wire configuration. This was seen in calculated models, as well as in normal specimens of the thoracic spine that were obtained from cadavera. Little epidural space was found to be available for passage of the wire. In most instances, passage of the wire must result in contact with and displacement of the dural sac and its contents. To minimize the depth of penetration at any given spinal level, it is recommended that the wire be curved to the maximum degree that will allow it to pass under the lamina.(ABSTRACT TRUNCATED AT 250 WORDS)