The unit of selection in Drosophila mercatorum. I. The interation of selection and meiosis in parthenogenetic strains

Abstract

An important problem in population genetics is the determination of the level of genetic organization to which fitness measures can be ascribed that yield an adequate description or prediction of the outcome of selection in populations. To study this problem, we used two strains of Drosophila merca torum (S-1-Im and O-3-Im) that are capable of both sexual and parthenogenetic reproduction, a feature that allows us to experimentally control many factors which affect genetic variability. Both S and O reproduce parthenogenetically by "pronuclear duplication," a mechanism that retains normal meiosis (and hence crossing over and assortment) but results in homozygosity for all loci in a single generation. Since an isozyme survey indicated that S and O differ at a third of their loci, we hypothesized that S and O have adapted in genetically distinct fashions to the genetic environment of total homozygosity. This is tested by breeding females that are S-O hybrids for 100%, 60% and 40% of their genetic backgrounds, and scoring their respective pathenogenetic progenies for four isozyme and two visible markers. The data collected gave evidence for a coadaptation to total homozygosity involving non-additive and non-multiplicative interactions between non-alleles. As the perturbation of the parental coadapted genotypes by meiosis increases (i.e., the greater the degree of S-O hybridity), the level of genetic material which behaves as an additive/multiplicative fitness unit becomes larger. Selective neutrality of genetic variation may be an artifact of our failure to measure the proper genetic unit of selection.