Most experimental spinal cord injury studies described to date have relied on a limited number of injury gradations, and have tacitly assumed that outcome (functional, histological, and/or neurophysiological) is a monotonically graded function of injury severity. In contrast, the present study provides evidence that functional and morphological outcome after spinal cord compression injury may occur in a discontinuous, non-graded manner in response to linearly graded injury levels. The thoracic spinal cord of adult rats was transiently compressed to thicknesses from 1.8 to 0.8 mm in 0.2 mm steps, or sham injury was administered. Open field motor behavior and segmental reflexes were evaluated up to 21 days post injury and correlated with histological measures and injury level. The highest correlation was between histological outcome and open field motor scores. Among the six injury groups, only three significantly different outcomes were apparent in the open field, reflex, and histological measures, consisting of the injury group pairs 1.8/1.6, 1.4/1.2, and 1.0/0.8 mm. At day 21, the 1.8/1.6 mm injury groups were also indistinguishable from the sham injury group. The implications of these findings in terms of therapeutic studies are discussed. Comparison of the temporal outcome patterns among contusion and compression injuries in rats and other species also revealed a significant species difference: a period of delayed or secondary functional loss reported in the guinea pig was not present in the rat.