Assessment of stature has been a neglected part of the medical assessment of individuals with spinal cord injury (SCI). In past studies of pulmonary function in SCI, it has not been stated how stature was assessed in order to calculate predicted pulmonary function. As part of a study of respiratory function in SCI, we examined the accuracy of self-report of stature and whether upper extremity measurements could be used to predict stature in 88 individuals with SCI. Although armspan and four other upper extremity measurements were significant predictors of length, recalled height was the best predictor. On average, for any value of armspan, an individual with complete SCI was 2.3 cm shorter than an individual with incomplete SCI. Individuals with complete SCI also had greater differences between recalled height and measured length compared with individuals with incomplete SCI, and this difference was not due to age or years since injury. It is likely that differences in bone demineralization account for the differences in length comparing individuals with complete and incomplete SCI. The 95 percent confidence intervals for predicted values of length based on armspan or recalled height were too wide for accurate calculation of predicted pulmonary function. In order to classify accurately the extent of pulmonary function abnormality, we suggest that measurements of supine length be made part of the medical assessment of individuals with SCI.