EVOLUTIONARY MECHANISMS 283 



Here we must discuss how a single population can become separated 

 into two groups which follow divergent evolutionary paths. 



If we have a homogeneous population of a species living under 

 uniform conditions in a certain area, natural selection will be working 

 equally on all parts of the species, and although it may cause a gradual 

 transformation of the species so as to adapt it better to its environment, 

 there, is no reason why the population should break up into two or more 

 separate specific groups. We should expect to find the gradual trans- 

 formation of the entire population described in the Gryphaeas and 

 Micrasters. Distinct species can only be formed if the hereditary 

 material of the population is split into two or more parts which do not 

 mingle. This splitting will occur if the population is divided, by geo- 

 graphical or ecological barriers, into groups of organisms which do not 

 interbreed. 



If we examine pairs of nearly related species which have probably 

 diverged from each other only recently it is rather commonly found 

 that some of the gene differences which accumulate in the two races 

 are such as to lead to complete infertility of the hybrids, even if this 

 was not present originally. For example, in Drosophila pseudoobscura 

 there are two local races known as race A and race B. These differ in 

 several specific inversions, but the chromosomes are all able to pair 

 with their homologues of the other race; the hybrid males are never- 

 theless sterile, and this is apparently due to gene differences. Dobz- 

 hansky^ has shown that numerous genes are involved, located in all 

 the chromosomes. It is interesting to find that within each race there 

 exist different strains with slightly different complexes of these "steril- 

 ity" genes; it seems that the interracial sterility is built up from a basis 

 of the minor differences occurring between the strains. Similarly, the 

 sterility of hybrids between D. melanogaster and D. simulans is not 

 caused by a failure of meiosis conditioned by chromosome changes, but 

 by gene differences which prevent the germ cells even beginning to 

 undergo the reduction divisions.- Other pairs of species may be pre- 

 vented from interbreeding by ecological factors, such as different dates 

 of maturing, or lack of sexual attraction, etc. 



Now the genes responsible for interspecific infertiHty or failure to 

 breed could not, in general, spread within an interbreeding population 

 and cause it to split into two groups, since they would lower the average 

 fertility, and thus the selective advantage of the organisms which 



^ Dobzhansky 19366. Cf. Dobzhansky and Koller 1938. 

 ^ Kerkis 1933. 



