172 MIMICRY IN N. AMERICAN BUTTERFLIES 



The geographical distribution of Imlsti strongly 

 supports the conclusion that it was derived from 

 archippus and not immediately from an arthemis-like 

 ancestor. I have not yet had the opportunity of 

 ascertaining whether this hypothesis is supported 

 by evidence derived from a careful study of the 

 pattern. 



It is deeply interesting to observe that the same 

 Limenitis arthemis-likQ species, from which archip- 

 pus, floridensis and hulsti mimics respectively 

 of the three Danaidas, pkxippus, Berenice and 

 strigosa have been directly or indirectly evolved, 

 has also given rise to L. astyanax (ursula), the mimic 

 of a Papilionine model. Evidence in favour of the 

 comparatively recent origin of these mimicking 

 forms is to be found in the well-supported facts 

 which indicate that astyanax still interbreeds with 

 arthemis along their geographical overlap, and 

 that it may even occasionally pair with the sister 

 species archippus. 1 



The earlier stages of archippus and astyanax are, 

 according to Scudder (l.c., 254, 255), with difficulty 

 distinguished from those of artliemis, but astyanax 

 presents the closer likeness of the two ; a fact 

 which, together with those referred to -in the last 

 paragraph, points to the conclusion that it arose 

 even more recently than archippus. 



The further consideration of astyanax is best 

 deferred until some account has been given of the 



1 Scudder, 1. c., 283, 289. Trans. Ent. Soc. Lond. (1908), 473, 474. 



