1909] Mimicry in the Butterflies of North America 221 



ment. Before the growth of a mimetic likeness, Batesian 

 mimics, it is reasonable to assume, belonged to the immense 

 group of species possessing a cryptic appearance; Miillerian 

 mimics on the other hand may be assumed to have possessed 

 warning or aposematic colours of their own previous to the 

 adoption of those of another species. This test is more readily 

 applied than might be supposed; for a comparison with allied 

 non-mimetic species, and with the non-mimetic males of mimetic 

 females, will generally indicate whether the ancestral pattern of 

 a species now mimetic belonged to the group of concealing colours 

 or to that of warning. 



The Danaidas invaded North America and entered an assem- 

 blage of butterflies of which the dominant species are ancient 

 inhabitants of the northern land-belt. Among them are sev- 

 eral such as the species of Grapta or Polygonia (the "Comma" 

 butterflies) , with beautifully cryptic patterns on the parts of the 

 wing surface exposed in the resting position. No such forms 

 have been influenced by the invaders, but w^ith the whole fauna 

 before them they have only produced changes in the dominant 

 group Limenitis, known throughout the northern belt for a con- 

 spicuous under surface and a floating flight; also believed to be 

 mimicked by other butterflies, e. g., the females of the Apaturas 

 ("Purple Emperors") and the later brood of Araschnia levana.'^^ 

 Furthermore, the close allies of Limenitis in South America, the 

 abundant Adelphas, are beautifully mimicked, not only by 

 females of the genus Chlorippe, which represents Apatura, but 

 also by Erycinidae. In another point the facts are at variance 

 with Bates' interpretation but harmonize with Miiller's. Bates 

 supposed Mimicry to be an adaptation by which a scarce, hard- 

 pressed form is enabled to hold its own in the struggle for exis- 

 tence But L. arthcmis, which represents with little or no change 

 the species from which the mimics were derived, persists as a very 

 abundant and flourishing species, while its mimetic descendant 

 archippus has gained an immensely extended range and become 

 almost universally commoner than any other species of its group 

 (Scudder, 1. c. 266). L. archippus extends from Hudson's Bay 

 to the Gulf of Mexico; over this vast area it is only rare in the 

 west, and only unknown in Colorado, Arizona, and New Mexico 

 (1. c. 278). It is to be observed that the range of archippus in- 

 cludes the whole of the area (Canada and the north-eastern States) 

 occupied by the ancestral form arthcmis. 



-^ See also tlie mimetic resemblance to L. astyanax described on pp. 229-30. 



