Offhand, I do not know of how we can gauge unreported incidents that would be more than big estimates. We have a comparison of what was voluntarily reported before November 1. We now have some indication of what's been reported since November 1.
Have we studied this issue...?