Don't forget you don't have to have the counts being correct -- you just want the comparison to be in the ideal ballpark, and when there's an get of magnitude variance, as there is here, It truly is pretty good proof of prevalent utilization, which is usually what you need to know when m