Neurohistologic examination of the spinal cord and cauda equina were compared for 28 beagles undergoing anterior and posterior spinal destabilization procedures--Group I (n = 7), destabilized operative controls; Group II (n = 7), posterolateral bone grafting; Group III (n = 7), Harrington distraction instrumentation and posterolateral fusion; and Group IV (n = 7), Luque rectangular instrumentation and posterolateral fusion. All dogs had appeared neurologically intact upon repeated examinations prior to death. Neurohistological abnormalities (Wallerian degeneration of the dorsal columns, corticospinal tracts, and nerve roots, focal cystic degeneration, and intraspinal central cavitation) occurred in only 1 of the 14 animals (7%) in Groups I and II (noninstrumented) and in 9 of the 14 animals (64%) in Groups III and IV (instrumented). This result is statistically significant (p less than 0.001). Transient sensory disturbances and radicular paresthesias have been described in clinical reports of spinal instrumentation. It is probable that subclinical neurologic injuries, such as intraspinal and nerve root infarction in posterior neural tissue, can occur with the use of sublaminar hooks or wires. The chondrodystrophic beagle spinal model in this study should be considered a "worst case situation," and the clinical incidence of neurohistologic changes is expected to be lower.