322 INTRODUCTION TO EVOLUTION 



which of two subspecies is represented. One reason for this state of affairs is 

 the fact that, as stressed earlier in our discussion of human races (pp. 251- 

 254), subspecies frequently differ from each other in the frequencies with 

 which certain genes (and hence characters) occur rather than in possession 

 of certain genes by all members of one subspecies and absence of those genes 

 from all members of a second subspecies. Varying distributions of the blood 

 group genes in different human races furnish an example (pp. 251-252). 

 "Races may be defined as populations which differ in the frequencies of 

 some genes" (Dobzhansky, 1950). 



Frequently the differences between subspecies seem to be more or less 

 chance differences of no particular significance in the lives of the organisms 

 (nonadaptive traits). At times we may consider differences nonadaptive 

 merely through ignorance of their real significance, but there is no reason 

 to doubt that populations may come to possess characteristics which are ac- 

 tually nonadaptive (see discussion of genetic drift, pp. 439-447). On the 

 other hand, differences between subspecies may at times be clearly adaptive. 

 A most interesting case of this in Peromyscus is afforded by a race of mice 

 inhabiting Santa Rosa Island off the coast of northern Florida, near Pensa- 

 cola. This long, narrow island running parallel to the shore and separated 

 from it in places by only a quarter of a mile of water is covered with an 

 exceptionally white sand. Living on the island is the lightest-colored race of 

 Peromyscus known: P. polionotus leucocephalus. "In this race, most of the 

 hair is white from base to tip, while the pigment of the skin is greatly re- 

 duced" (Sumner, 1932). Such a white race living in regions of white sand 

 affords an excellent example of protective coloration, the protection in 

 this case being against predators that use their sense of sight in locating 

 prey, especially owls. We shall describe later (pp. 364-365) experiments 

 which demonstrated that in actuality being light colored on a light-colored 

 background does protect mice from predation by owls. We may picture this 

 white race as having arisen during the past few thousand years by the action 

 of natural selection favoring genes and mutations tending to lighten the 

 color of the mice inhabiting the island. We may note in passing that leuco- 

 cephalus is not an albino, differing from normally pigmented mice by a 

 single gene as albinos commonly do. From the results of breeding experi- 

 ments Sumner (1932) concluded that leucocephalus differs from the fully 

 pigmented polionotus (see below) by a number of genes having additive 

 effects (multiple genes or polygenes, pp. 391-394). 



The race leucocephalus was evidently derived from the somewhat similar 

 race on the neighboring mainland: Peromyscus polionotus albifrons. This 

 race lives on the beaches and is light in color but not so light as leuco- 



