Heredity (1981), 47 (3), 439-442 0018-067X/81/03770439$02.OO 1981. The Genetical Society of Great Britain SPECIATION AND INFERENCES ON RATES OF MOLECULAR EVOLUTION FROM GENETIC DISTANCES ALAN IR. TEMPLETON, ROB DE SALLE and VIRGINIA WALBOT Department of Biology, Washington University, St. Louis, MO 63130, U.S.A. and Department of Biological Sciences, Stanford University, Stanford, CA 94305, U.S.A. Received 12.v.81 MOST descriptions of molecular evolution ignore the initial genetic distance associated with the process of speciation (Templeton, 1980a) and can therefore lead to an oversimplification or unnecessary exclusion of possibilities. For example, Cianchi eta! (1980) have recently demonstrated that Nei's genetic distance between pheromone strains of the European corn borer is 0004 for enzyme coding loci classified as "non-regulatory" by the scheme given in Johnson (1974), whereas the distance is 0056 for loci coding variable substrate or regulatory enzymes. One interpretation of these data is that different rates of molecular evolution occur in these different enzymes classes (Cianchi et a!., 1980). It is the purpose of this note to point out that although this interpretation is consistent with the data, alternative interpretations exist as well. According to Nei (1975) the genetic distance between two species that split into independent lineages t time units ago is: D=Do+2at (1) where a is the rate of molecular evolution and D0 is the initial genetic distance between the lineages. It is commonplace to assume that D0 = 0 (Nei, 1975) so that D = 2at. Hence, any difference in genetic distance between classes of gene loci would have to be attributed to differences in a, the rate of evolution. However, the assumption that D0 = 0 effectively assumes that the establishment of lineages is an instantaneous event in time with no genetic implications of its own. No evolutionary justification for this assumption has been explicitly given, and indeed none exists (Templeton, 1980a). When speciation is regarded as a process rather than an event it becomes apparent that the assumption D0 = 0 cannot be made a priori (Templeton, 1980a). In general, D0 will be some positive number that depends upon the mode of speciation, the population genetic con- straints affecting that mode of speciation, and the level of polymorphism for the loci being used to measure genetic distance. For example, one mode of speciation is the genetic transilience (Templeton, 1980b) in which a founder event directly induces the erection of isolating barriers. The initial genetic distance between founder and ancestor is given by (Templeton, 1980a) D0 = (1— G)/(4NG) (2) where N is the number of founders and G is the average homozygosity in the ancestors under Hardy—Weinberg expectations. As the level of poly- morphism increases, G tends to decrease and hence the initial genetic distance tends to increase. This is true for other modes of speciation as 439