THE ORIGIN OF SPECIES 



identical for each. Also, no distinction was made between different species 

 of birds. Now it is hardly to be supposed that any animal would be per- 

 fectly protected from all of its potential enemies; but if an animal is even 

 somewhat protected from some of its potential enemies, it ought to have 

 an improved chance for survival in a highly competitive animal society. 



There is some positive evidence that presumed protected animals are 

 not taken by predators as often as are unprotected species. Examples have 

 been given above for aposematic animals. Again, the protection need not 

 be absolute. Thus, the porcupine is not preyed upon by the vast majority 

 of carnivores, but the fisher (Martes) successfully takes porcupines by 

 turning them over on their backs, then ripping open the unprotected 

 ventral surface. But one would hardly wish to claim that, because a single 

 carnivore takes porcupines, their protection is not effective. For there are 

 many carnivores which would otherwise kill porcupines. 



Experimental evidence is also available in support of the above observa- 

 tions. Sumner's experiments on predation on the mosquito fish (Gambusia) 

 by penguins are most illustrative. These fishes slowly change color to 

 match their background. If one group of fish is kept in a black tank and 

 another in a light tank, each group will become adapted to its own back- 

 ground. If fishes from one of the tanks are then transferred to the other, 

 they will contrast with their background until they can again become 

 adapted. Sumner exposed such mixed groups of adapted and unadapted 

 fishes to predation by penguins. Always, both types of fish were taken, but 

 the unadapted fishes were taken in much greater proportion than were the 

 adapted ones ( Table 3 ) . The differences are highly significant statistically 

 in both types of tank, and they thus support the protective value of the 

 color changes. 



Table 3. Predation by Penguins on Adapted and Nonadapted Mosquito Fish * 



* By permission from Cott, "Adaptive Coloration in Animals," Methuen & Co., Ltd., 

 London, 1940. 



Dice exposed variously colored races of the white-footed mouse, 

 Peromijscus, to predation by owls. When several color races were exposed 

 on a background nearly matching one of them, that one was taken less 

 frequently tlian were the others. 



The extensive literature dealing with all phases of adaptive coloration 

 has been summarized and exhaustively studied by Cott. He has concluded 

 that, although the concepts involved have been much damaged by too 



248 



