J. A. MOORE 335 



postulate two populations, C and D, which were isolated for some 

 time during which they diverged genetically to such an extent 

 that any hybrids they might form would show reduced viability. 

 Let us further assume that C and D reestablish contact and in 

 the zone of overlap they hybridize. Since we have postulated 

 that the hybrids have reduced viability, let us assume with Dob- 

 zhansky, that natural selection would promote the development 

 of isolating mechanisms. These isolating mechanisms would have 

 selective advantage in the zone of overlap hut not elsewhere. 

 Consequently they could not become characteristic of the species 

 as a whole, for the following reason: If we assume with Dob- 

 zhansky that "Isolating mechanisms encountered in nature appear 

 to be ad hoc contrivances which prevent the exchange of genes 

 between nascent species," and that they are formed by selection 

 in zones of overlap, we can assume that they have adaptive value 

 only in the zone of overlap. If they were of adaptive value to 

 the two species in areas where they were not sympatric, they 

 would be selected but not for their function as isolating mech- 

 anisms. Genetic differences that are ad hoc isolating mechanisms 

 between two species can be selected for only in areas where the 

 two species occur sympatrically. These isolating mechanisms can 

 develop as a consequence of selection for genes that prevent 

 the formation of hybrids or by selection for genes that prevent 

 competition ( irrespective of whether or not hybrids are formed ) . 



Outside of the zone of overlap any gene differences that serve 

 as ad hoc isolating mechanisms would probably be either neutral 

 or somewhat deleterious. (If they were advantageous they would 

 be selected for any way, but with no reference to any pleiotropic 

 functions as isolating mechanisms.) It is difficult to understand, 

 therefore, how they could become established as isolating mechan- 

 isms throughout the ranges of both species unless the ranges were 

 coextensive. Except on small islands, this last possibility would 

 be highly unlikely. 



Several additional reasons can be cited for believing that the 

 origin of isolating mechanisms by selecting against hybrid forma- 

 tion cannot be of universal occurrence. 



First, there are some species pairs that are able to form appar- 



