Indirect negative evidence for frequency-dependent selection with respect to viability in Drosophila melanogaster

Using the estimates of additive genetic variance and dominance variance as well as the regression coefficient of heterozygotes on the sums of the component homozygotes (βY-x) and the inverse of the regression coefficient of the sum of the two homozygotes on their heterozygotes (l/fβY-x} with respect to viability estimated by the Cy method, it was tested whether or not frequency-dependent selection is a general rule with respect to viability polygenes. Two types of frequency-dependent selection models were considered: (1) gene frequency-dependent model and (2) zygotic frequency-dependent model. The βy.x and 1/βx-y values predicted on the basis of the frequency-dependent selection models and the ratio of the dominance variance to the additive variance are inconsistent with the actual estimates of these parameters for the Florida, Ishigaki-jima (Japan), Raleigh (N.C.) and Aomori (Japan) populations. The same oonclusion was obtained for newly arisen polygenic viability mutations. Thus, it is speculated that frequency dependent selection is not a general rule for viability polygenes as well as polymorphic isozyme genes.