THE SPHINGO-MICKOPTERYGIDES. 121 



the larva may be accompanied by a generalisation of the moth." It is 

 notorious, Dyar adds, that the Eucleid larvae live for an unusually 

 long period. 



The Megalopygid larva has the arrangement of the tubercles the 

 same as in the Pyromorphid (i.e., Anthrocerid). It also has two 

 additional pairs of prolegs, viz., those on the 2nd and 7th abdominal 

 segments (paralleled only perhaps by traces of extra pairs in Pselno- 

 phorus br achy dacty his on the 1st and 2nd abdominal segments). 

 These two additional pairs were carefully figured and described 

 (Surinamsche Vlinders, I., pi. xiv., and explanation) by Sepp. With 

 Packard's conclusion " that this larva represents, as no other known 

 caterpillar the polypodous ancestor of all Lepidoptera," we are quite 

 unable to agree, since, in our opinion, the abdominal prolegs are quite 

 secondary structures. The condition of the rudimentary prolegs on 

 the 2nd and 7th abdominal segments shows, we think, the manner in 

 which the sucking discs of the Eucleid larvte have been developed, and 

 how the latter have obtained their peculiar mode of progression. 



As we have already pointed out, the range of larval characters in 

 the Pterophorids is a wide one, and shows how modification may be 

 brought about in an otherwise closely related superfamily. In the 

 miners, the tubercles bear single seta?, and are degenerate. In others, 

 the tubercles are modified into warts that are largely characteristic of 

 the generalised superfamilies of this stirps. In Aydtstis, the" dorsal 

 tubercles of the 9th abdominal segment are modified into a caudal 

 horn. Dyar, speaking of the relationship of the Anthrocerid and 

 Pterophorid larvae, says° : "The Anthrocerid larva, A. filvpendulae, has 

 the tubercles converted into warts, but the warts are greatly reduced, 

 being represented by tufts of short hairs. Tubercle if is absent, iv 

 and v are consolidated, and vii is very distinct on the base of the leg." 

 Here, Dyar says, we have direct evidence of the continuity of this 

 series of families with the ancestors of the Pterophorids and their 

 allies. Further, Dyar states that the Pyromorphids (Anthrocerids) 

 show essentially the same structure as the Megalopygid and Eucleid 

 larva?, but there are no urticating spines, and the prolegs are normal. 



There are two distinct lines in which specialisation of the tubercles 

 of the larvaa of this stirps, produces a very similar result, although the 

 essential characters of the development are very different. Thus 

 specialisation here takes place in the reduction of the number of 

 tubercles, and this may occur : (1) By the union of two or more 

 primitive tubercles, as in the union of iv and v to form a single sub- 

 spiracular wart in the Anthrocerids. (2) By actual disappearance of 

 the tubercles as in some Eucleids, Bombycids, Saturniids, etc.J It is 

 quite clear that it is possible, therefore, to have a similar general result 

 arrived at by two entirely different processes of evolution, and to have 



*" Classification of Lepidopterous LarvsE," Ann. Neiv York Acad. Set., viii., p. 211. 



t On the contrary, tubercle i is coalesced with ii, forming a large dorsal wart, in 

 which there is a distinct trace of the two separate portions of which it is formed. 

 See our description of this and the allied larva, later in this book. 



I There is a tendency, in Lasiocampid, Bombycid, and Saturniid larvaa, 

 very slight in some species, but very strong in others, for tubercle i to become 

 greatly enlarged and specialised, and for ii to become atrophied. In the early 

 stages of Saturnia imvonia and Bombyx mori, ii exists as a minute tubercle with 

 (or without) a single seta, whilst i is large and many-haired. At a subsequent 

 moult ii is lost by atrophy, it does not merge into i. 



