1909] Mimicry in the Butterflies of North America 227 



utely large numbers of such specimens with those from districts 

 where the model is dominant. A similar study should be made 

 of the Canadian specimens of asterius, although this species does 

 not extend so far beyond the northern limits of the poison-eating 

 model. 



From another point of view the interbreeding of the turnus 

 female of glaiiciis with a male from some northern district where 

 turnus is unknown or very scarce would be of the highest interest. 

 We should here be able to test whether the Mendelian relationship 

 exists between the parent form and its partially melanic variety 

 further transformed by selection, — not a mere melanic "muta- 

 tion." I trust that my friend Prof. C. B. Davenport may be 

 able to undertake this experiment at the Cold Spring Experi- 

 mental Station. I cannot doubt that breeding could be easily 

 carried through two generations in a large enclosed space exposed 

 to the sun and planted with abundant flowers and the food-plant 

 of the species. It would probably be safe to use Long Island 

 males, while female pupae or the freshly bred females themselves 

 could be readily obtained from further south. 



The Evolution of Limenitis (B.) astyanax (F). as a Mimic 

 OF Ph. philenor and Its Papilio Mimics. 



Scudder states that L. astyanax "ranges from the Atlantic 

 westward to the Mississippi Valley, and from the Gulf of Mexico 

 northward to about the 43rd parallel of latitude."^* It thus falls 

 entirely within the area of philenor. The northern boundary 

 of astyanax corresponds with the southern limit of its parent 

 arthemis, and Scudder (1. c. 289) considers that they interbreed 

 and that the intermediate form proserpina, found along the nar- 

 row belt where the two species or sub-species meet, is the result- 

 ing hybrid. Both arthemis and proserpina have been bred from 

 the eggs of the latter. There seems little doubt that astyanax is 

 a very recent development from arthemis in the southern part of 

 its range, — so recent that the areas of distribution still remain 

 distinct and parent and offspring only meet along a narrow line. 

 It is probable that archippus arose in the same manner in part 

 of the area of arthemis, but that later, after the separation had 

 become complete, it spread northward over the whole range of 

 its parent. 



-^ A closelj^ allied species or probably a form of the same species is recorded 

 bv Godman and Salvin from Mexico. 



