According to this little article in the Annals of Improbable Research, Kansas is flatter than a pancake. I dunno. Seems to be a bounding problem, since the edges of the pancake are included. Since Kansas has a politically, rather than naturally set boundary, I think the border values of the pancake should have been excluded. That is, the edges should not be trimmed off, an action which physcially depresses the pancake, but mathematically removed from the analysis. This introduces a different problem – how far in should you cut? Since Kansas is part of the American plains, should you trim off until the area of the pancake specimen equals the relative area that Kansas occupies in the Western Plain?
Inquiring minds would like to know. Or maybe not.
And with that, we return to our regularly scheduled programming. (via Volokh)