CHIEF CONCLUSIONS SUMMED UP 205 



between the larval and pupal stages of parent and 

 offspring. L. archippus also probably occasionally 

 interbreeds with the mimetic L. astyanax — a still 

 younger descendant of the same parent. 



9. L. arcMppus probably arose on the southern 

 borders of arthemis, but afterwards ranged north- 

 wards over the area of the parent species. 



10. The southern astyanax, meeting the northern 

 arthemis along a narrow belt, is probably repeating 

 the earlier history of archippus. 



11. The forms or sub-species of archippus — 

 floridensis in Florida and hulsti in Arizona — have 

 arisen from the earlier mimic of D. plexippus as 

 a result of the predominance ia these localities, re- 

 spectively, of Banaida lerenice and its form strigosa. 



12. Details of the older Mimicry persist in 

 floridensis (and perhaps in hulsti), somewhat de- 

 tracting from the newer resemblance. 



18. Certain features in the mimetic likeness 

 newly attained in Florida and Arizona are prob- 

 ably due to the recall or the re-emphasis of 

 elements in the pattern of arthemis which had 

 been greatly reduced in archippus. 



14. The factthat the invading Danaidas haveonly 

 influenced, among the whole indigenous butterfly 

 fauna, the dominant conspicuous Nymphaline 

 genus Lim^nitis, supports a Miillerian as opposed 

 to a Batesian interpretation of the phenomena. 



15. The fact that the ancestral pattern of a 

 species indigenous in the temperate zone of the 

 New World should be wholly transformed by 



