ISOLATION 155 



1926, p. 40). While it is possible to assume that these forms 

 evolved in geographical isolation but have since crossed their 

 barriers, it is doubtful whether the evidence for the necessity 

 of geographical isolation is so cogent that it is necessary to 

 make so big an assumption. 



We may summarise the outstanding points in this con- 

 troversy as follows : 



1. The male armature differs specifically much more 



often and usually more markedly than the female. 



2. There is often, perhaps usually, no close specific corre- 



lation between the male and female structures. At 

 least such correlation has not been established. 



3. The numerous interspecific crosses, mostly artificial but 



some natural, between species with very different geni- 

 talia, show that the male and female armatures by no 

 means necessarily impose an insuperable barrier. 



4. The vast mass of species with different genitalia prob- 



ably do not try to interbreed. They are in fact 

 separated by other types or combinations of types 

 of isolatory factors (especially those included under 

 I and II (a) ). 



5. As a corollary to (4), large groups of species exist in 



which the female genitalia differ but little from 

 species to species. There is no evidence that such 

 forms hybridise more readily than those in which 

 the differences are marked. 



6. There appears to be no very high correlation between 



degree of differences in genitalia and the fertility of 

 hybrids if a pairing does take place — e.g. Sturtevant 

 (1920) shows that Drosophila simulans and D. melano- 

 gaster have identical mating habits and hybridise 

 freely, but the hybrids are quite sterile. The male 

 genitalia differ, but not those of the females. 



We can only conclude that the genital armature may 

 sometimes provide a bar to interspecific crosses, but the bar is 

 by no means universal or incapable of being surmounted. 

 This is particularly true of the smaller differences which 

 characterise very closely allied species. The value of specific 

 differences in the genitalia lies rather in their relative 

 constancy. Thus, while variation does occur {e.g. marked 



