.Inalysis of the Theory of ll'aniiiig Coloration. 315 



rests leaves one impressed with its soundness. The resemblance between 

 mimicked and mimickcr have been characterized by Bates as at times "per- 

 fectly staggering." Wallace states (1891, p. 245) that "in almost every 

 box of butterflies brought from tropical America by amateurs are to be 

 found some species of the mimicking Pieridse, Erycinid3e, or moths, and the 

 mimicked Heliconidse, placed together under the impression that they are 

 the same species. Yet more extraordinary, it sometimes deceives the very 

 insects themselves." The differences between mimicked and mimickers may 

 afford us a measure then of the power of discrimination of insectivorous 

 vertebrates, for according to the theory those mimics which differed notice- 

 ably from their models have been continually destroyed by their vertebrate 

 foes until the survivors have come to be so like their models that they can no 

 longer be distinguished from them. Any greater differences must attract 

 the attention of these foes. Insectivorous vertebrates thus show a power of 

 discrimination exceeding that of many amateur collectors. 



That the gray snapper possesses some such power of discrimination 

 may be inferred from the experiments already described. Bateson says that 

 the wrasse " finds a shrimp if the least bit exposed, in spite of its protect- 

 ive coloration" (Poulton, 1890, p. 204). If this power be conceded to 

 fishes, insectivorous reptiles, birds, and mammals may well possess at least 

 equal nicety of discrimination. Recognition marks, if they are recogni- 

 tion marks, afford another instance of discriminative power among verte- 

 brates, while protective and aggressive resemblance are quite generally ac- 

 cepted as having arisen through selection which has depended on the powers 

 of discrimination of vertebrate foes. 



The theory of warning colors requires us, on the other hand, to assume 

 that these same insectivorous vertebrates possess so little power of discrimi- 

 nation that unpalatable insects need to be marked in conspicuous patterns of 

 contrasting colors. May we not with more reason assert that, if insectivor- 

 ous vertebrates have pushed the resemblance between mimickers and their 

 models to the point of apparent identity, warning colors are for them quite 

 unnecessary? The ordinary specific differences should suffice to "warn" 

 them of the unpalatability of prospective and familiar insect prey. If this 

 be true warning coloration has no more title to be so called than have other 

 distinctive characters of insects. That it actually serves to " warn " insec- 

 tivorous vertebrates can not be doubted, but that it is necessary to that end 

 may be seriously questioned. 



With reference to tenet II (b) above. If insectivorous vertebrates pos- 

 sess the nicety of discrimination required by the theories of mimicry and 

 protective resemblance, then it is difficult to see how selection can have 

 produced warning coloration. Selection can no more account for incipient 

 warning coloration than it can account for other incipient characters, for 

 like other characters, warning coloration must have developed to the point 



