The neuropathology of 20 cases of traumatic human spinal injury is described. Cases were classified as having ''complete'' or ''incomplete'' spinal cord injuries based upon the neuropathological findings. Special attention was directed to those cases with neuropathological evidence of continuity of long tract axons across the lesion but that were obtained from individuals with clinically complete sensory and motor loss. Clinical and neuropathological data were obtained from the Western Australian Spinal Cord Injury Database and were selected from those cases involving sports-related injuries. From the total of 20 cases of spinal injury, 6 exhibited neuropathologically and clinically complete cord injuries, 4 were neuropathologically incomplete yet clinically complete, and 4 were neuropathologically and clinically incomplete (cone being motor complete yet sensory incomplete). The clinical/neuropathological correlation was indeterminate in 6 cases because of death at or shortly after the accident. The neuropathological findings varied with the severity of trauma and the length of survival. In those cases examined less than 3 months postinjury, there was typically swelling of the cord from vasogenic edema with petechial hemorrhage and myelomalacia. Central hemorrhagic necrosis was common as was hyperemia with free red cells present among disrupted tissue. Activated astrocytes and polymorphs or macrophages were observed and early stages of glial cell scar formation were evident. Wallerian degeneration was evident in motor tracts caudal to the level of insult and in sensory tracts rostral to the lesion. Cases examined 3-6 months postinjury often revealed multilocular cysts with gliotic walls often extending several millimeters above and below the primary locus of injury. In the ''chronic'' cases (i.e., those surviving more than 6 months), there was evidence of varying degrees of preserved long tract parenchyma, nerve root regeneration, and Schwann cell remyelination of long tract axons. The degree of sparing of long motor and sensory tract parenchyma generally correlated well with the clinical observations of residual function. In four cases, however, there was a residuum of motor or sensory tract axons that appeared to have been spared injury, but the clinical data reported chronically complete sensory and motor loss. The presence of clinically complete yet neuropathologically incomplete injury was more often associated with crush or flexion injuries than with extension or compression trauma. The preserved parenchyma in these types of lesion may provide the anatomical substrate for restorative intervention.