A Classification of Lepidoi')terous Larvae. 195 



based on the larval tubercles might differ in important respects from 

 one based on the adult insect. That this is not the case in the 

 present instance for the major groups, shows that the classification 

 proposed by Prof. Comstock is the nearest to a natural one that we 

 have yet had. That it is frequently the case in the minor groups, 

 illustrates the principle to which I have just referred. It is, more- 

 over, a fact, that we see the greatest amount of modification in those 

 larvae which live the most exposed life, and whose period of larval 

 existence is the most extended. Some of these cases will be speci- 

 ally alluded to. 



General External Structure of Lepidopterous Larv^. 



The body of the lepidopterous larva is divided into thirteen seg- 

 ments, three thoracic and ten abdominal.^ The last two abdominal 

 segments are closely united in most of the higher families, so to 

 appear as one. On each of the third to sixth and tenth abdominal 

 segments are a pair of fleshy legs, furnished with prehensile hooks. 

 In special cases, certain of these legs are absent or modified for other 

 uses ; or, much more rarely, the three pair of jointed, chitinous, 

 thoracic legs may be wanting. On each segment (when not too 

 highly specialized) are certain, small, chitinous tubercles, arranged 

 in a definite manner, each bearing a bristle or seta. The arrange- 

 ment and modification of these tubercles is the subject of this paper. 

 On the thoracic segments, the arrangement is different from that 

 on the abdominal ones. For simplicity, we will confine attention 

 principally to the latter, and to those segments of them which bear 

 feet. 



Since all lepidopterous larva? periodically cast their skins (vary- 

 ing from three to ten times in different species) before the final molt 

 to acquire the pupa stage, and since the external structure is not the 

 same in each larval stage, it is necessary to consider at what stage 

 the larval tubercles are to be compared. There are two types 

 among the more highly specialized larvae. In the first, represented 

 by the Saturniidse and Notodontidte, the arrangement of tubercles 



1 Dr. A. T. Bruce remarks in the case of the embryonic Thyridopteryx ephe- 

 tneraifonnis : "It is questionable whether the terminal portion of tlie abdomen 

 which forms the so-called eleventh abdominal somite is to be regarded as a 

 true somite or not. It has no ganglion corresponding to it, and is formed . . . 

 by the dorsal flexure of the posterior end of the embryo. ..." (Observa- 

 tions on the embryology of insects and arachnids, p. 9, 1887.) 



