GENE FREQUENCY IN SMALL POPULATIONS OF MUTANT DROSOPHILA

Studies of gene frequency changes in laboratory populations have for the most part been concerned with the systematic effects of selection and with discovery of the nature of the selective differences involved between genetic alternatives. The classical experiments of L'Heritier and Teissier (1934, 1937) with common mutants, the later work of Dobzhansky (Wright and Dobzhansky 1946), and that of Wallace (1948) on naturally occurring genetic variants are examples of such studies applied to rather large populations maintained in population cages of the sort developed by L'Heritier and Teissier (1933). Such experiments as these were carried out under conditions in which changes in gene frequency due to sampling accidents (random drift) were negligible. In the work of Reed and Reed (1948, 1950), Ludwin (1951) and Merrell (1953) on selective differences between mutants and wild type, non-systematic effects of considerable magnitude and attributable to accidents of sampling occurred; however, these data did not permit a close evaluation of the extent of this influence. An experimental design in which both random and systematic pressures are brought into play in an orderly fashion and the values of the several theoretical parameters can be ascertained was originated by Kerr (Kerr and Wright, 1954). The method consists in observations on the changes in gene frequency between generations in each of a large number of small cultures in which the size of the breeding population is controlled. Each initial culture founds a line in which each successive generation is inititated with a random sample of uniform size taken from among the flies of the preceding generation. The form is that of a fairly large experimental population divided into a number of very small completely isolated subunits. The present study was intended to provide information on the relative selective values of two mutant autosomal alleles and their several zygotic types under several different experimental conditions and to observe the effect of the interaction of selection and random drift on the distribution of gene frequencies among the elements of subdivided populations of the sort observed by Kerr.