When less is more: founder flush

A population bottleneck has a tendency of reducing the long term effective population size (harmonic mean) a great deal. I'm sure all of you knew that, and I'm sure you know that it also has a tendency of reducing variation. This is because low effective population sizes increase the power of genetic drift, which tends to expunge variation from a population. Great. The only thing though is that sometimes bottlenecks can "release" variation and make it available for selection. Additive genetic variance (heritability) increases as the genetic background is fixed on many loci, but not all, via a founder flush. I point this out just to suggest that sometimes the inferences drawn from population genetics can go against our expectation, because expectation in a probabilistic science must be modulated by variance.1 - The approximate time until fixation of a neutral, that is totally dependent on drift, mutation is 1 X 4 X effective population. So, if you crank down the effective population new mutants will fix very quickly in populations and you have a situation where transient polymorphism is reduced.