122 BEITISH LEPIDOPTEBA. 



two superficially similar types having no real phylogenetic relationship. 

 It is, therefore, important to compare the results obtained from the 

 imago, pupa and egg with those obtained from the larva, otherwise, 

 one may readily fall into error. Thus the actual relationships and 

 position of the generalised members of this stirps, the Anthrocerids 

 (including the Pyromorphids), the Eucleids and the Megalopygids, 

 are very doubtful if the characters of the larvaa alone be considered. 



We have already dealt with the relationships exhibited by the 

 generalised superfamilies of this group, and we have already pointed 

 out, that one of the most important larval characters in them is the 

 union of tubercles iv and v into a large sub-spiracular, many-haired 

 wart, after the first larval stage. Dyar notices that this generalised 

 character is also found in otherwise highly specialised Saturniids, and 

 is strongly suggestive of the alliance of the Saturniids with this group. 

 The Lasiocampid and Bombycid larva3 also show many characters that 

 ally them with the more generalised superfamilies of the stirps, 

 although in the former the development of a hairy coat has led to the 

 obliteration of the tubercles, iv being, indeed, almost obsolete. There 

 seems to be but little difference between the urticating spines of the 

 hairy Eucleids and those of the Lasiocampids, Packard stating that 

 those of Empretia stimulea are only loosely attached, as is the case 

 with those of Lasiocampa qaercm and Macrothylacia rubi, and we look 

 upon the Lasiocampids as the most nearly allied of the specialised 

 families of this stirps, to the generalised Eucleids and Megalopygids. 

 The break between the Lasiocampids and Megalopygids is, how- 

 ever, a very wide one. On the other hand, we understand that 

 the Megalopygid egg is very like the Anthrocerid (not so scale-like as 

 the Eucleid) egg ; but, in spite of this, the Megalopygids are still 

 closer to the Eucleids than to the Anthrocerids, the generalised pupae 

 and the neuration being very similar, and so far as the larvae are con- 

 cerned we are inclined to look upon the creeping discs of the Eucleid 

 larva as exhibiting a specialised form derived from the extra pro- 

 legged Megalopygid larva, for, in the latter, the abdominal segments 

 2 and 7 bear extra prolegs without hooks, and are not very dissimilar 

 from what might be assumed as a first stage in the development of 

 the Eucleid suckers. At any rate, so close is the alliance, and so far 

 back in the evolutionary period are the Eucleids, that at the time of 

 their origin there must have been great plasticity as to prolegs, and 

 the specialisation is not difficult to understand. The entire absenc© of 

 armature in the highest Eucleid larvae tends to show that the Eucleid 

 larva is essentially a specialised form of the hairy Megalopygid. It is, 

 therefore, from an ancestor resembling the former rather than the 

 latter, that we are inclined to derive the Lasiocampids, and we find 

 that the latter have retained certain generalised characters exhibited 

 by the former, of which the habit of covering the eggs with silken 

 hairs, the possession of specialised, urticating, larval hairs, the 

 peculiar '' eggar " cocoon, with its separately formed lid, the thin 

 transparent pupal integument, and the pectinated antennae of the 

 imago are still common to both. Cs For these reasons we are inclined 



* Some of these peculiarities are, we know, to be found in species belonging to 

 other stirpes, and it is quite possible that some of them, at least, were derived from 

 an ancestor even older than the Megalopygid, from which both the Megalopygids 

 and the Liparids obtained their special peculiarities in this direction. 



