We don't have that information readily available now. We can certainly look back. We do know from past studies that typically we're talking 5% of collisions being attributable to maybe a vehicle defect. It's a small portion. I don't have an exact number to provide the committee.
On March 11th, 2010. See this statement in context.