The biomechanics of indirect reduction of bone fragments retropulsed into the spinal canal in a burst fracture were investigated. In this model, tunnels were created in vertebrae L1 and C5 oriented anterior-to-posterior, allowing access to the posterior longitudinal ligament. A probe containing a load-sensing tip was passed through the tunnel. Both the location of the tip and the load acting on it by posterior deflection of the posterior longitudinal ligament were measured. In the lumbar spine, distraction was applied by spinal instrumentation that also permitted independent kyphotic-lordotic alignment of the vertebrae. In the cervical spine, axial traction was applied through direct loading. Several clinically relevant observations were made. It was not possible to produce an anteriorly directed force in the posterior longitudinal ligament at less than 35% canal occlusion, partly because the posterior longitudinal ligament stands away from the midbody of the vertebra. Distractive forces of up to 150 N were applied in the lumbar spine, which were nearly equal to the tensile breaking strength of the isolated posterior longitudinal ligament. Regardless of the relative sagittal plane angulation of the vertebrae, distraction was the governing factor in generating force in the posterior longitudinal ligament. Because positioning the vertebrae in lordosis before applying distraction significantly slackens the posterior longitudinal ligament, it is suggested that distraction be applied before angular positioning of the vertebrae is performed.