152 THE CANADIAN ENTOMOLOGIST. 



color of the aphides on alder that I should not suppose the wool would be 

 necessary to their preservation." 



Also : " I have never found a chrysalis, though I have looked for 

 them whenever I have been in the swamps. My larvae generally seek a 

 leaf, but I think it probable the wild ones crawl down the stems and 

 pupate among sticks or grass." The larva of Lye. Fseudargiolus when 

 ready to pupate drops to the ground. 



The observations settle these points : that the eggs are laid directly 

 among the aphides, and in case of stem-aphides, on the bark ; that the 

 ants do not destroy the eggs (though usually ants destroy every egg they 

 find) ; that the larvae from egg to past second moult conceal themselves 

 under the aphides, and under spun webs of loose texture, through the 

 meshes of which they devour their prey, and which webs also serve to 

 protect them from injury, especially at moulting time ; that the larger 

 larvae, that is, from before third moult on, are in full view, but besides 

 being coated with wool from the aphides, have ways of protecting them- 

 selves from enemies, as by falling off the stem, throwing out a thread, or 

 by falling to the ground ; that there is no period, at any moult, of much 

 length when the larva is helpless, and apparently none at all at third 

 moult, when it is most exposed; that there are but three moults, and the 

 whole larval period is exceptionally short ; that the larvae will eat many 

 species of aphides (possibly any, unless deterred by certain species of 

 ants), but prefer the large, woolly ones. 



I have repeatedly had letters from different parts of the U. S. and 

 Canada, asking if I knew on what the larva of Tarquinius fed, and in 

 nearly all cases the writers stated that the butterfly had been taken on or 

 near alders. 



Prof Riley, in the Science paper quoted, notes that this is the only 

 butterfly known whose larva is carnivorous. But next to nothing is known 

 of the early stages of tropical butterflies, especially in the great family of 

 Erycinidae. Both there and among the Lycaenidae there may be species 

 which have this same peculiarity. 



Godart conjectured that Tar-quinius should be classed with Erycina. 

 His instinct was right ; Fenesica belongs to the Erycinidae. The present 

 classification of butterflies, based as it is solely upon one stage of the 

 four, is imperfect and at best but temporary, and is sure to give way to a 

 better as the early stages of species become known. 



At Coalburgh there would appear to be at least three broods of the 



