NATURAL SELECTION IN ACTION 365 



so, mice having mutations (or other genetic variability) making for light 

 coat color would be expected to survive on the island in larger proportion 

 than would mice having darker coloration. Thus the lighter-colored mice 

 would contribute a larger proportion of genes to the next generation than 

 would the darker-colored mice. If this trend continued for many genera- 

 tions, the present almost white race of mice could be accounted for. We 

 note that in this explanation predators utilizing vision for hunting occupy 

 a key position. Granted that light coat color causes the mouse to blend 

 with the light background to our eyes, does it also do this in the eyes of 

 predators? Does "protective coloration" actually protect? 



In order to answer this question Dice (1947) performed a series of ex- 

 periments with Peromyscus differing in shade, utilizing owls as predators. 

 A darkroom was divided into halves by a low partition on one side of 

 which light-colored soil covered the floor, on the other side dark-colored 

 soil. Both dark-colored and light-colored mice were placed in both halves 

 of the room. The owl lived in a nest box near the ceiling in the middle of 

 the room. Would the owl catch more dark-colored than light-colored 

 mice on the light soil, and more light-colored than dark-colored ones on 

 the dark soil? At first there seemed to be no tendency of this kind. Grad- 

 ually the experimenter reduced the dim light intensity used during the 

 tests until finally there was no light at all. Still the owls caught mice. Evi- 

 dently, then, they were not using the sense of sight in their hunting. Prob- 

 ably they located the mice by hearing their movements. Marks on the soil 

 indicated that in the darkness they used their wings to sweep in mice 

 located by hearing. 



The problem was, then, to force the owls to rely on the sense of sight. 

 The experimenter did this by covering the floor of the room with an arti- 

 ficial "jungle," a sort of latticework of light timber arranged so that owls 

 could reach through the meshes and catch mice when the light intensity 

 was sufficiently high so that the mice could be seen. This "jungle" simu- 

 lated the plants and bushes under which mice normally live. In a series 

 of trials under these conditions 107 conspicuous mice (dark-colored on light 

 soil, and light-colored on dark soil) were captured but only 65 concealingly 

 colored ones were. Dice found that "in every experiment in which the 

 predator was evidently using sight to capture his prey, the conceahngly col- 

 ored individuals enjoyed more than a 20 percent advantage over the con- 

 spicuous animals in escaping capture." And he concluded: "such a high 

 rate of selection, should it be applied to a natural population, would un- 

 doubtedly result in a very rapid change in the frequencies of the genes 

 producing the character under selection." 



