770 THE BUTTERFLIES OF NEW ENGLAND. 



forming a family group of similar significance. The structures upon 

 which one may draw for the characterization and delimitation of this 

 group will be discovered in every stage of life. The egg differs in a very 

 marked manner from that of any other group Jiowever small among but- 

 terflies (with the single exception of the Parnassians), in being univer- 

 sally broader than high, oblate spheroidal, and generally turban-shaped. 

 The an-reement of the two subfamilies, Lemoniinae and Lycaeninae, in 

 the characteristics at this stage of life are so complete that it is imijossible 

 with our slight knowledge of the former, to formulate any satisfactory 

 diagnostic distinctions; and they wholly differ from the egg of the 

 Parnassians in the character of the surface sculpture, the cells of the 

 reticulation being lower than the bounding walls, while in the Pai-nas- 

 sians the reverse is the case, giving them the appearance of being covered 

 with plates. 



The caterpillar as it leaves the egg is distinguished by the presence 

 of chitinous annuli or lenticular elevations, serially arranged on the 

 dorsal side of the body ; these are not found at this stage, at least in 

 the same region, in any other group, and they are here found in both of 

 the subfamilies. The adult caterpillar is remarkable for its generally 

 onisciform shape, so that in the Lycaeninae at least they are recognizable 

 at a glance ; but even where they are cylindrical they differ from those of 

 every other group in their abbi-eviatcd form. In the other subfamily, the 

 Lemoniinae, there is greater variability of form, but we never find any with 

 the great elongation of the body characteristic of all the other fiimilies. 

 In both they differ from most others and agree together in the incomplete 

 structure of the posterior part of the head, the chitinous skull presenting 

 here no downward slope, the softer membrane of the succeeding seg- 

 ment being attached to the skull near the summit of the head, so that the 

 head becomes more or less retractile, sometimes certainly to a very slight 

 degree, within the first thoracic segment, while the head itself has a very 

 decided obliquity. These last characters, however, they share to a certain 

 extent with the family Papilionidae. 



The chiysaliils of the two groups agree with each other again and differ 

 from those of all others in their compactness and brevity, while at the 

 same time they offer one peculiarity, found in no other group and which 

 holds here throughout both; viz., the head cannot be seen from above, 

 being bent over and forming a part of the ventral surface only ; sometimes 

 the same is true of the last abdominal segment. The prothorax, too, is 

 proportionately larger than in any other group, and these ^peculiarities of 

 the head and prothorax the two subfamilies hold in common. Add to this 

 the character long known to be peculiar to them, the close girding of the 

 chrysalis, with the flatness and uniformity of the ventral surface, characters 

 which with rare exceptions they share together and in which they differ 



