The current study was undertaken to follow the time course of bone loss in the proximal tibia of rats over several weeks following thoracic contusion spinal cord injury (SCI) of varying severity. It was hypothesized that bone loss would be more pronounced in the more severely injured animals, and that hindlimb weight bearing would help prevent bone loss. Twenty-six female Sprague-Dawley rats (200-225 g, 6-7 weeks old) received standard thoracic (T9) injuries at energies of 6.25, 12.5, 25, or 50 g-cm. The rats were scored weekly for hindlimb function during locomotion. At 0, 2 or 3, and 8 weeks, high-resolution micro-CT images of each right tibia were obtained. Mechanical indentation testing was done to measure the compressive strength of the cancellous bone structure. The 6.25 g-cm group showed near normal locomotion, the 12.5 and 25 g-cm groups showed the ability to frequently or occasionally generate weight-supported plantar steps, respectively, and the 50 g-cm group showed only movement without weight-supported plantar stepping. The 6.25, 12.5 and 25 g-cm groups remained at the same level of bone volume fraction (cancBV/TV=0.24±0.07), while the 50 g-cm group experienced severe bone loss (67%), resulting in significantly lower (p<0.05) bone volume fraction (cancBV/TV=0.11±0.05) at 8 weeks. Proximal tibia cancellous bone strength was reduced by approximately 50% in these severely injured rats. Instead of a linear proportionality between injury severity and bone loss, there appears to be a distinct functional threshold, marked by occasional weight-supported stepping, above which bone loss does not occur.