PLEIOTROPY AND PARAPATRIC SPECIATION.

Abstract

Parapatric speciation as described by Murray (1972), Bush (1975), and Endler (1977) is the process of species formation from two or more populations that initially have some gene flow among them. This process is different from allopatric speciation, which requires that there first be a complete absence of gene flow, and sympatric speciation, which requires that there initially be random mating (Bush, 1975). Parapatric speciation was discussed in genetical terms by Fisher (1958, p. 142-143), who postulated that if populations were subject to different selection pressures in different locations, then natural selection would reduce the gene flow among them. Fisher proposed two mechanisms for reducing the gene flow, lower dispersal tendencies and greater selectivity of the females in each population for males from their own population. modification of dispersal has not received further discussion as a mechanism of speciation because it does not lead to reproductive isolation, but the evolution of assortative mating-now called the (Murray, 1972)-has been the principal mechanism in existing models of sympatric and parapatric speciation (e.g., Maynard Smith, 1966; Dickinson and Antonovics, 1973; Endler, 1977; Felsenstein, 1981). A consequence of the Wallace effect is that the resultant species are isolated only by pre-zygotic isolating mechanisms. Another way for reproductive isolation to evolve is as a byproduct of genetic changes at selected loci. Muller (1940, 1942) and Thoday and Gibson (1970) used the information on pleiotropy in Drosophila to argue that speciation could result from such pleiotropic effects. Wright (1969 and elsewhere) has long emphasized the importance of pleiotropic effects of most mutations, and Mayr (1963, p. 551) implicitly invokes pleiotropy in the establishment of reproductive isolation in allopatric populations: The ecological shifts in incipient species are bound to have an effect on their isolating mechanisms. thesis of the origin of reproductive isolation as a byproduct of the total genetic reconstitution of the speciating population is consistent with all known facts. In this paper I will investigate the potential importance of pleiotropy in parapatric populations and show, by means of a simple model, that pleiotropic effects at different loci tend to reinforce each other and can lead to speciation in parapatric populations. Reproductive isolation can be a byproduct of genetic changes at a locus under two conditions. Either the locus could control two characters, one of which is subject to natural selection and the other of which affects mating success, or the locus could control a single character which serves two purposes. Traditionally, a pleiotropic locus is one that affects two or more characters, and this is the kind of pleiotropy that Muller and others have had in mind when discussing the role of pleiotropy in speciation. Maynard Smith (1966) argued that this type of pleiotropy was unlikely to be of importance for speciation. However, the second possibility, namely a character that is strongly selected also affecting mating success, may well be important for speciation. A good example of a character that serves two roles is flowering time in plants. McNeilly and Antonovics (1968, Table 2) list 28 examples of known ecotypic differences in flowering time. Differences in flowering time could evolve completely through the Wallace effect as a mechanism causing assortative mating between ecotypes. McNeilly and Antonovics (1968) provide strong evidence