Yes I have those. I've since called Atmel and quess what... they don't have a clue either. His first suggestion was to replace my RC reset circuit with just a cap, but I'm positive that this won't do it. There is just too much code at the beginning that works as it should, before this variable get initialized. It has to have something to do with the ISP boot. I'm just getting ready to eliminate the resistor to satisfy Atmel