Because S doesn't vary linearly in that interval, a+1/2 isn't the expected population for a state somewhere in that interval.
If I didn't have things that need to be done right now, I'd write that expected population correctly in this posting. As it is, though, I have things to do, and must log off. Milke Ossipoff
---- Election-Methods mailing list - see http://electorama.com/em for list info
