During traumatic brain and spinal cord injury, central nervous system (CNS) tissue undergoes mechanical deformations at rates sufficient to cause a range of deficits. In order to determine the effect of strain and strain rate on a nerve bundle — and on axons of different diameters within that bundle -- a series of experiments were performed which exposed isolated crayfish ventral spinal cords to a range of strains and strain rates believed to be responsible for CNS injury. Suction electrodes were used to monitor the electrophysiological behavior of the cord, and indicators of injury were significant changes in area under the compound action potential (CAP) compared to preinjury data, and significant changes in heights of peaks in CAP curves. Seven groups of excised crayfish cords, with an average of five cords in each group, were tested for 25 minutes before and 25 minutes after imposed injury. Levels of strains and strain rates in the five non-control groups ranged from 10% quasi-static elongation to 30% strain in approximately 10 ms. Results demonstrate that (a) the magnitude and rate of tensile loading of in vitro nervous tissue dictate the severity and duration of the resulting dysfunction, and (b) smaller axons sustain higher immediate and prolonged injury than larger axons. A mathematical model was also developed to investigate the longitudinal strain experienced by sinusoidal axons when a measurable deformation is applied to the matrix in which they reside. A number of researchers have reported on the sinusoidal orientation of axons in a nerve when the nerve is observed at its in vivo length. This is protective of the axons, because if tissue including the nerve bundle is elongated, the axons will not experience the same strain as the entire tissue. Results from two submodels indicate that the protective effect of geometry is very significant: Sinusoidal axons may experience 20-80% lower longitudinal strain and strain rates than the surrounding matrix depending on the loading conditions. Further, for moderate levels of axonal sinusoidal curvature, matrix elongations of up to 22% may not actually lengthen axons whatsoever.