Heterosis in crosses between geographically separated populations of Drosophila melanogaster

Summary An experiment was performed to test the hypothesis that the genetic distance between populations estimated from enzyme loci could be used to predict the amount of heterosis that would occur in crosses between these populations. A partial diallel cross using 11 populations of Drosophila melanogaster from the AustralianPacific region and from England was carried out. Heterosis for larval viability, fecundity, cold shock mortality, and an index of these three traits was recorded. When two populations originating from the same location were crossed, no heterosis occurred, but otherwise heterosis was significant for all traits. For larval viability, a similar low level of heterosis occurred in all crosses. For cold shock mortality, the level of heterosis varied widely and fecundity showed a pattern intermediate between these two. The geographic distance between the sites from which populations originated was not correlated with the amount of heterosis in their crosses. There was a tendency for populations from ecologically different environments to show heterosis in crosses. Genetic distance based on ten enzyme loci was correlated with heterosis for cold shock mortality and the combined trait index. These results can be explained by the hypothesis that genes affecting larval viability are subject to strong, uniform selection in all populations, which limits the extent to which gene frequencies can drift apart. However, genes affecting cold shock mortality and the enzyme loci are subject to different selection pressures in different environments. This divergent selection combined with genetic drift causes divergence in gene frequency and heterosis.