NATURAL SELECTION 



651 



number of concealed mice compared to the 

 number of conspicuous mice taken by the 

 barn owl (T. alba pratincola) in two ex- 

 periments were 68:124 and 65:107. The 

 selection indices were 0.292 and 0.244, 

 respectively, while the chi-squares were 

 16.333 and 10.256, indicating high statisti- 

 cal significance. Dice concludes that such 

 high selection of subspecies variation in 

 nature would produce rapid evolution to- 

 ward protective coloration (see pp. 610, 

 627, 668; Figs. 230, 245). 



as having warning (aposematic) or con- 

 cealing (procryptic) coloration, forty-five 

 were offered, and thirty-eight, or 84 per 

 cent, were accepted. Of the somewhat more 

 conspicuous insects, thirty-five were oflFered, 

 and two, or 6 per cent, were accepted. Of 

 the typically conspicuous insects with warn- 

 ing coloration, fifty-eight were offered, and 

 five, or 9 per cent, were accepted. The spe- 

 cies differences in these tests are presumed 

 to be genetic in the majority of cases on the 

 basis of their taxonomic correlations. These 



Fig.. 237. A praying mantis resembling the ckad brown lea\es among which it li\es in Panama. 



(Photograph by Ralph Buchsbaum.) 



Carrick (1936) tested a number of spe- 

 cies of insects against insectivorous birds 

 that were feeding their young. He classified 

 the specific degree of resemblance to the 

 surroundings in five categories. Of the typi- 

 cally concealed insects, hidden by their ex- 

 treme degree of resemblance to the colors 

 (and often shapes) of their immediate sur- 

 roundings (see Figs. 236 and 237), forty- 

 three were oflFered, and seven, or 16 

 per cent, were accepted. Of the insects 

 with dingy colors generally resembling the 

 surroundings, sixty-three v/ere oflFered, and 

 fifty-three, or 84 per cent, were accepted. 

 Of the neutral insects not readily classified 



experiments are interesting in their demon- 

 stration of relative freedom from predatory 

 attack of both the inconspicuous and the 

 conspicuous forms, thus lending credence 

 to the concept of background (cryptic) 

 coloration (Figs. 236, 237 and 244), warn- 

 ing coloration (Fig. 238), and by inference 

 from warning coloration, to the theories of 

 mimicry (p. 670). It should be noted that, 

 in Carrick's experiments, no diflFerential sur- 

 vival value is demonstrated between those 

 insects with a general resemblance to their 

 background and those classified as neutral. 

 Turning from selection through preda- 

 tors (Errington, 1946) to the selective ef- 



