I have performed an AMOVA analysis in pegas as well as GenAlEx (and gotten the same results, phew). I have two regions and several populations in each region. In both pieces of software, I get a table with three rows: among regions, among populations within regions, and within populations. Theoretically I could divide "among populations within regions" into "among populations within region 1" and "among populations within region 2" but I don't know if there is any software to do this. With the output of GenAlEx it is easy to partition the sums of squares, but I am not sure how I would calculate the variance given that the calculation of N0 is a bit beyond me (or, I have tried using formulae that I have found online, but have not gotten the right answer to reproduce the results).