182 Increase in Number of Species 



and clearings, and E. fiaviventris chiefly in coniferous growth 

 (Peterson, 1947). Other examples involve differences in seasonal 

 timing or host. Epling (Dobzhansky, 1941) pointed out that certain 

 related species of Salvia flower at different periods, and Prison 

 ( 1935 ) found that certain related species of stoneflies had different 

 seasonal times of emergence. In both plant and animal parasites 

 which are restricted in host requirements, it is commonplace for 

 two closely related species to have different obligate hosts. The 

 different requirements of the individual species must be based on 

 associated physiological differences of some kind. 



These and innumerable other examples of closely related species 

 occupying the same geographic range but exhibiting ecological 

 segregation have suggested to many students of evolution that a 

 species can divide into two distinct genetic populations by ecological 

 isolation without the aid of geographic isolation. The basic reason- 

 ing behind this suggestion presupposes that if a species occurred 

 in two quite different ecological situations, the individuals in one 

 situation would tend to interbreed as a unit, as would those in the 

 other situation. In each situation different selection pressures would 

 operate on the pattern of genetic change so that the two popula- 

 tions would move apart genetically. The tendency of each ecological 

 population to remain separate would prevent genetic mixing be- 

 tween them. If this situation persisted, theoretically the two popula- 

 tions would become sufficiently different genetically that members 

 of one could not interbreed successfully with the other. At this 

 point two species would have evolved, each restricted and adapted 

 to distinctive ecological requirements. 



Except under peculiar circumstances involving either host speci- 

 ficity, seasonal timing, or unusual ecological gradients, there seems 

 to be no good reason to suppose that the set of suppositions above 

 would work. The chief drawback is the lack of any apparent reason 

 why, within the large network comprising the range of a species, 

 individuals from one situation would not mingle with and mate 

 with those from another, unless the opposite ecological conditions 

 were situated some distance apart. In this case genetic divergence 

 of the populations would be a function of partial or complete 

 geographical isolation, not ecological isolation (Mayr, 1942; Lack, 

 1944). Thus Blair (1947) reports finding no dificrences in gene 

 frequencies between mice on differently colored soils only four 

 miles apart but marked differences between populations on the 

 same colored soils 18 miles apart. 



Ecological isolation has been invoked especially to explain 



