STILL ANOTHER ATTEMPT TO ACHIEVE ASSORTATIVE MATING BY DISRUPTIVE SELECTION IN DROSOPHILA.

Abstract

Response to disruptive selection (Mather, 1955), either for stable polymorphism within a breeding population or for splitting of a population's gene pool into reproductively isolated separate gene pools, has been demonstrated experimentally in the past (Thoday, 1972). Which outcome might ensue is determined by dependency between disruptively selected genotypes, the persistence and intensity of selection applied, as well as the nature of the genetic variation available in the population. With current interest in the genetic basis of speciation, we have reconsidered the free mating choice experiments of Thoday (1963) and Thoday and Gibson (1962, 1970) with Drosophila melanogaster that had been selected disruptively for a morphological feature (stemopleural chaeta number). Since these investigators were able to obtain significant reproductive isolation within a coherent gene pool, others (including ourselves here described) have attempted to repeat Thoday's experiments but have failed to achieve significant reproductive isolation (Scharloo et al., 1967; Chabora, 1968; Barker and Cummins, 1969a, 1969b). The basic design was to apply disruptive selection for stemopleural bristle number within a genetically variable strain of D. melanogaster with the aim of producing two mutually isolated subdivisions. Both with the Thoday experiments and with the repeats by the other investigators cited in 1967-1969, the technique involved letting matings take place among several pairs in which the bristle number was either or and then discarding the males without regard for the specific male member of each mated pair. Progenies counted with respect to the female parent were scored and then selected in turn. Actual matings had not been observed, but in Thoday's experiments the fact that the two controls of hybrids between high and low strains produced intermediate numbers of bristles in progenies indicated the main cause for divergence to be assortative mating. We postulated that selection of mated pairs for the sum of their joint bristle numbers would be more efficient than previous experiments in obtaining contributions of both sexes toward the morphological trait and thus toward any assortative preference, provided that genetic variation was available for preferential mating. We expected this selection scheme to exploit any genetic variation in the mate recognition system (in Templeton's use of the term, 1979), selecting directly for positive assortative behavior if sternopleural bristle number has any role in the behavioral system. This system, a coadapted signal-response chain of the male and female activities that facilitates mating, must depend on discriminatory ability of females predominantly. While sternopleural bristle number cannot be considered a trait neutral to selection (Kearsey and Barnes, 1970), its role in mating behavior is nevertheless speculative; certainly bristles have sensory receptors at their bases (probably tactile), but we do not know whether the association between this morphological feature and reproductive behavior is one of pleiotropy of a few genes, cytogenetic linkage of loci controlling these traits, or some indirect physiological relationship as yet unknown. There is no question that some association exists in certain strains, since Thoday obtained rapid isolation using the