MEMOIKS OF THE NATIONAL ACADEMY OF SCIENCES. 53 



of Lcpidoptera, only the larvre of the Tiicboptera, Pauorpid;?, and Tenthredinida; approaching- 

 them, they do not seem to aflbrd salient features of value for subordinal characters. Yet there 

 are some archaic features, such as the arrangement of the hooks on the abdominal legs, the 

 presence of cvcrsibk' coxal glands on the under side or on the sides of the body; and in the larva 

 of Eriocephala we have subordinal characters in the absence of a functional spinneret, also in the 

 extraordinarily large size of the antenna^, and of the maxillary palpi of that genus. 



The process of specialization in the larva has effected not so much the general form of the 

 body as the armature of the abdominal legs and of the body. Chambers, and also Dinimock, 

 (Psyche, ill, 99, 1880) has shown in Lithocolletis and in Gracilaria, especially, the changes which 

 take i)lace in the head and mouth parts as well as feet of the larva after the first molt, in adapta- 

 tion from a mining to a free existence. But in free-feeding forms it is difficult to distinguish a 

 normal Tineid larva from a Tortricid or Pyralid larva, and as yet no characters diagnostic of them 

 and other I'amilics have been indicated. With the exce]itiou of the larv;e of certain Tineina, of 

 the Cochliopodida' (Limacodid;e), of the Psychida', those of the Hesperians and the onisciform 

 caterpillars of Lycaenida-, lepidopterous larvaj are remarkably homogeneous in form, as they are 

 in habits. The only reliable larval characters for distinguishing families are the differences in the 

 piliferous tubercles, the number of hairs or seta' arising from a tubercle, or the shape and size of 

 the tubercles themselves, and even within the limits of anj- family there is great variation in 

 these, as seen in the Saturniida', or the Ceratocampidaj, or Arctiidw, etc. 



The resemblance between the larva' of the Trichoptera and the Lcpidoptera is remarkably 

 close, their internal and external anatomy being nearly the same, the Lcpidoptera differing chietiy 

 in the presence of abdominal legs; these, however, being absent in ^licropteryx. 



Supposing that tiie Lcpidoptera did spring from some neuropterous group allied to the stem 

 form of the Trichoptera, the type at once after the primitive lcpidoptera ceased to live in the water, 

 if its ancestors were aquatic, assumed abdominal legs, hooks developed on them, at first a pair, then 

 more until two complete rows appeared, and the larva was fitted to climb the stems of plants in 

 order to feed on the leaves. Eventually Ave may imagine that the larva', owing to the attacks of 

 insect parasites, sought shelter by mining leaves, seeds, twigs, stems, trunks, and even roots of 

 ]>lants. In adaptation to these novel surroundings, the mining forms by disuse lost their legs, 

 their bodies became flattened and otherwise modified as in certain Tineina, or the sack bearers 

 were modified in adaptation to their peculiar habits. Tliis great diversity in the mode of obtain- 

 ing their vegetable food and their exposure to varying surroundings resulted in manifold special 

 adaptations in ornamentation and armature, hence the groups most successful in the struggle for 

 existence became very numerous in genera and species. 



The generalized forms may be detected by the larvie having one-haired warts, with minute 

 tubercles without spines, but other primitive forms have large tubercles, warts, humps, or highly 

 colored lines, bands, or spots. While the larval characters are useful in distinguishing genera 

 or families, they do not appear to present salient subordinal characters, as they do in Coleoptera, 

 Diptera, and Hymenoptera. 



0. The generalized pupal forms are those nearest to the pupa libera of Trichoptera and the 

 Neuroptera, etc.; such is that of ^licroptcryx. Those pupa^ with more or less free abdominal 

 segments, the Pnpw incompletce of Chapman, are plainly more archaic or generalized than those 

 belonging to his division, Pujxe obtecta', which comprise the modern or specialized forms. Where 

 the ends of the maxillary palpi appear externally under the eyes; where the labial palpi are 

 visible; where what we call the paraclypeal pieces are present, we have survivals of the characters 

 of the pupa libera of Micropteryx. When these features have been by modification lost, we have 

 the uniform obtected pupa of the Neolepidoptera, and these characters are so persistent that they 

 are of high taxonomic value. 



7. The pupa, then, is of the greatest importance in defining the larger groups of the haustellate 

 Lepidoptera, and chiefly for the reason that the lepidopterous pupa, with its so-called wing and 

 appendage cases, appears ti) represent not only wliat maybe called a subimaginal condition, but a 

 still earlier, lost, or extinct imaginal type, a type perhaps midway between the ametabolous and 

 metabolous series. This is suggested by the wing-cases which are as in ametabolous nymphs, 

 such as those of Dermaptera, Termitida-, and Psocida', as well as of Hemiptera; and, as shown 

 by Spuler, the venation of the lepidopterous i)upa is almost identical with that of the Blattidw 



