Analyzing variation in egg-to-adult viability in experimental populations of Drosophila melanogaster

Selective culling in populations of most organisms is rank-order: individuals of low rank on a scale of potential fitnesses tend to be eliminated during early development, whereas surviving adults (whose number reflects the carrying capacity of the environment) are generally drawn from the distribution's upper end. Haldane pointed out [Haldane, J.B.S. (1932) The Causes of Evolution (Harper & Row, New York)] that selection which favors individuals in the upper tail of a composite distribution curve tends to favor members of the more variable of two populations, rather than members of a less variable one, even though the latter may possess the higher mean. In addition to reviewing earlier observations bearing on Haldane's argument, the present report describes an analysis of the comparative egg-to-adult viabilities of flies (Drosophila melanogaster) carrying combinations of second chromosomes obtained from one or another of eight experimental populations. Overall, the viabilities of flies carrying combinations of chromosomes one of which is shared (i/j vs. j/k) are as different as those of flies carrying combinations of independently sampled chromosomes (i/j vs. k/l). Episodes seemingly occurred within the populations during which flies carrying combinations that shared a chromosome differed more in their viabilities than flies carrying unrelated combinations. Such episodes could reflect the occurrence of selection of the sort described by Haldane.