476 INTRODUCTION TO EVOLUTION 



course of a few generations. Koopman ascribed this decline to the action 

 of natural selection in the population cages since he concluded that the 

 poor viability and sterility of hybrids and their offspring would have 

 prevented their contributing to future generations even if he had not 

 removed them. (See also Merrell, 1953.) 



Similar results were obtained by Knight, Robertson, and Waddington 

 (1956) who employed ebony-bodied and vestigial-winged stocks of 

 Drosophila melanogaster. Males and females of both stocks were placed 

 together so that they could either mate with their own kind or cross-mate, 

 as they "wished." But in each generation only offspring of homogamic 

 (pure-breeding) matings were used as parents for the next generation. As 

 generations passed, production of hybrids declined, indicating that some 

 degree of sexual isolation between the stocks had been produced by 

 selection. 



Hence the results cited seem to indicate that under experimental condi- 

 tions, at least, selection will intensify one of the mechanisms of re- 

 productive isolation: selective mating. 



Moore (1957) has pointed out that factors increasing reproductive 

 isolation would be of value to the populations concerned only in those 

 regions where the populations were actually in contact. (No advantage to a 

 frog living in Vermont would accrue from having a genetic constitution 

 whose only function would be to render impossible hybridization with a 

 Florida frog with which it would never actually come into contact any- 

 way.) Blair (1955), investigating two species of frogs with overlapping 

 ranges, noted that the greatest difference in mating call and the greatest 

 difference in size between the two species occurred in the region where 

 their ranges overlapped. He concluded: "The existence of the greatest 

 size differences as well as the greatest call differences where the two 

 species are exposed to possible hybridization supports the argument that 

 these potential isolating mechanisms are being reenforced through natural 

 selection." 



Such reenforcement is not always encountered, however. Thus Volpe 

 (1955) found that reproductive isolation between two species of toads was 

 weaker in areas where the ranges overlapped than it was in other regions 

 (judging from the results of laboratory experiments on artificial hybridiza- 

 tion). He cited other investigations yielding similar results. 



Perhaps some of the conflicting results arise from differences in the 

 animals being investigated; production of hybrids is not detrimental to all 

 species and under all conditions (see below). Natural selection, if it does 

 act to favor development of isolating mechanisms, will only do so under 



