A significant lateral acceleration has been measured on vehicles tested in frontal offset car-to-car crashes. Typically, the lateral acceleration was -about 50% of the maximum longitudinal acceleration. The maximum longitudinal acceleration occurred during the period of maximum axial loading of the lower limbs of the test dummy. It has been postulated that the footwell acceleration may be a factor in ankle injury causation. This paper uses a finite element model of the lower limbs to explore the consequence of the transverse acceleration. The finite element model of the human leg was incorporated into an existing model of a Hybrid III dummy. The Hybrid III dummy has been previously validated, using sled and crash test data. The resulting dummy model permitted a comparison of the response of a Hybrid III leg , and a human leg. The dummy model was subjected to crash acceleration environments similar to those produced in offset crashes; however, no toepan intrusion was permitted. The response of the human leg model was much different from the dummy leg model. The dummy ankle rotated to the stops of 45 degrees in dorsiflexion, and +/- 20 degrees in inversion and eversion. The human ankle produced only 20 degrees of dorsiflexion, but 30 degrees of inversion/eversion. Limitations in the Hybrid III dummy biofidelity and injury criteria make inversion and eversion ankle injuries difficult to validate experimentally. The results with the human leg model suggest that inversion/eversion ankle injuries may be induced by transverse acceleration. The population with the lowest injury tolerance may be vulnerable to these injuries, even in the absence of footwell intrusion.