That additive genetic variance¹ can increase after a founder event because various alleles can fix onto the genetic background and subtract epistatic effects which would normally cancel out? This is important since response to selection ~ additive genetic variance * selection coefficient. Researcher dissents/critiques welcome.

Update: Consider epistatic effects which cancel out over the genetic background because the deviation is sometimes positive and negative (Fisher would be proud). A founder event could move some loci to fixation, which frees up some of the other loci² to work their additive variational magic. A rugged and fitness constrained landscape is now transformed into a clean and conical Fisherian geometry. Do you think this sort of process is evolutionarily relevant?

1 - The variance crucial for microevolutionary process, as in the breeder's equation.

2 - Which were previously hooked into pairwise fitness deviating epistatic interactions.