364 BRITISH LEPIDOfTERA. 



irritated. Dyar states that this larva seems to be the most specialised 

 form of the " spiny " type, and that those of four other species 

 examined exhibited a gradation from that of D. vulnerans, with 

 strongly eversible spines, to the smooth form. It appears probable to 

 us that the Australian forms are in this group, as in others, the oldest 

 we have, except that the hairy forms may be earlier as being related 

 to the Megalopygids (La(joa), which also may be earlier than tbe 

 Cochlidids (Eucleids). If this were so, the larva of our British species 

 Gocldidion (Limacodes) avellana, with its eversible spines on hatching, 

 would represent in its first skin, the adult larva of Doratifora. It 

 would also in its third skin, with subsidiary spines, represent the 

 latter forms with fixed spines, and give, as the most recently evolved 

 form, a smooth larva. If any forms become smooth in the second 

 stage they would, of course, on this supposition, be older still, but 

 at present there appear to be no individuals known with smoother 

 larvae than those of our British species. 



Dyar explains the structure of the Eucleid larva by reference 

 to the Anthrocerid type (Psyche, viii., pp. 171-174). Every British 

 lepidopterist will know that after the first moult the Anthrocerid 

 larva has, on either side of each segment, three complex warts, as 

 follows : (1) Subdorsal (formed from tubercles i and ii). (2) Supra- 

 spiracular (iii). (3) Subspiracular (ivand v). Dyar says that in the 

 Eucleid larva the subventral area is reduced, owing to the formation of 

 the creeping disc, and all the warts below the spiracular region are 

 obsolete. Assuming the Anthrocerid larva to have retained in this 

 particular the more primitive form, the extinction of the subspiracular 

 tubercle in the manner shown would leave the primitive Eucleid form 

 (as such) with three warts on each side of the thoracic, and two on 

 each side of the abdominal segments. 



The modification of these warts into their present forms is supposed, 

 by Dyar, to have taken place in two ways : (1) By hypertrophy, which 

 has resulted in producing the spined Eucleids. (2) By atrophy, which 

 has resulted in the smooth forms. On these lines, Dyar subdivides 

 the " hypertrophied " forms into two main larval " types " : (1) The 

 tropical hairy Eucleids (illustrated by Phobetroh and Calybia), with a 

 combination of general characters that suggests this as the most 

 generalised Eucleid type. (2) The tropical spined Eucleids (illus- 

 trated by Sibine, Euclea, etc.), to which many of the Indian and 

 South American species, figured by authors, belong. To this group 

 Dyar refers the Australian species, which differ from those of Asia 

 and America " in having the spines removed from the horns which 

 have not become eversible," a peculiar specialisation which leads Dyar 

 to consider the Australian type as the most modified of all Eucleid larva?. 

 The " atrophied " forms also, according to Dyar, fall into two main 

 divisions : (1) The tropical smooth Eucleids (illustrated by Eidima- 

 codes), which, starting with warts in the first stage, lose them by 

 degeneration after the first moult, their place being taken by single 

 setse. (2) The Palaearctic smooth Eucleids [illustrated by Gocldidion 

 (Apoda),Heterogenea and Packardia], in Avhich, Dyar says, the single seta? 

 of the first stage are retained and the warts have entirely disappeared, 

 leaving an almost smooth larva. To this last type, the European species 

 belong. Of course the spiny forms may have originated from a generalised 

 Anthrocerid-like larva, but, we think, there can be little doubt, after 



