120 BRITISH LEPIDOPTERA. 



The Lasiocampid larva, in many ways generalised, is specialised in 

 the direction of a thick hairy coat developed from the secondary or 

 skin (not tubercular) hairs, whilst the normal tubercles have become 

 much atrophied or flattened, although their position is similar to that 

 found in the more generalised superfamilies. The Bombycid and Eupte- 

 rotid larvre show affinities tending to lessen the distance between the 

 Lasiocampids and Sphingids, the former inclining to the Sphingids, 

 the latter to the Lasiocampids. The Endromid larva is remarkable 

 for its general Lasiocampid structure (warts, etc.) in the 1st skin, 

 and general Sphingid appearance in the later skins, whilst the 

 Sphingids and Saturniids present to us some of the most highly 

 specialised of lepidopterous larvae ; the former with its bright, 

 oblique, lateral stripes, and prominent caudal horn, the latter, with 

 its remarkable spines and hairs presenting to us larval types that 

 have lost almost all traces of the general characters that distin- 

 guish the least modified superfamilies. Chapman states that larvre 

 have not yet been found, that present characters to bridge over the dis- 

 tance between the Cochliopodids and Lasiocampids, but Packard hints 

 that the Megalopygid larva (Lirjoa) is connected with the Saturniids. 

 The observations, however, made by Chapman on the spines of the larva 

 of Apoda and Eadca, as well as on those of Sphingids and Saturniids, 

 tend to show that all are branches of the same stirps. The observa- 

 tions of Poulton and Weismann on the larvre of A<jlia and other 

 Saturniids, and the comparison of these with the Sphingid larvae, leave 

 little room for doubt that these superfamilies are somewhat closely 

 related. Packard's studies also indicate that simila.r relationships 

 exist, and he also is able to show considerable agreement between 

 the larvae of the Ceratocampids and the Sphingids. 



It may now be advisable to consider the larvae of some of the super- 

 families at greater length. In the larvre of the generalised members 

 of this stirps, specialisation (following the living of an exposed life) 

 takes place in different ways : (1) By the conversion of a simple 

 tubercle into a raised wart, the latter bearing many setae, resulting in 

 a hairy armature. (2) The reduction of the normal tubercles and 

 specialisation of colour for protective purposes. As to the reduction 

 of the tubercles, Dyar says : " This condition is to be seen in the 

 Pterophorids, where tubercle i has disappeared. Then there is 

 the reduction and disappearance of the subventral tubercles, leaving but 

 two above the spiracle and one below. This is probably in adaptation 

 to the form of the larva which is now becoming flattened with shortening 

 of the prolegs. The Pyromorphids and Megalopygids are in this 

 condition. Following this, the sub-stigmatal tubercles iv and v dis- 

 appear, and the flattening of the ventral region reaches its extreme. The 

 hairs of the tubercles have also been gradually converted into stinging 

 (? urticating) spines. In the Megalopygids, they are only partially con- 

 verted ; but in the Eucleids, the conversion is complete, and we have, 

 finally, the highest type of the Micro-lepidoptera (generalised), and, 

 perhaps, the most highly modified of all Lepidoptera. It is interesting 

 to note that the Eucleid moths are much more generalised than those 

 of several of the families which precede them, while the larvre 'are so 

 highly specialised. This illustrates the principle that specialisation in 



* " Classification of Lepidopterous Larvse," Ann. New York Acad. Sci., viii., p. 20C. 



