hmm. I would have thought this matching so well across random counties was impossible. well, one way to check this is to plot the same for another state's counties. if it is much different (not matching, all over the place) would you then think this graph was highly suspect?