58 TRANSACTIONS OF THE [Nov. 12, 
though no one seems to have adopted his conclusions.* Mr. 
Butler argues that the species of Acronycta should he distrib- 
uted among the Arctiide, Lymantriide, Notodontide and Noc- 
tuidz on account of larval resemblances. That is, though the 
moths are usually considered congeneric, the larve belong to 
several different families. This would mean, as the classifica- 
tion stood then, two superfamilies in the sense in which that 
term is used here. It is true the larvee of the genus Acronycta 
are wonderfully varied in appearance, but I believe that this di- 
versity is due to mimicry of all sorts of objects from that of 
resemblage to the foliage (grisea, tritona, etc.) to warning colors 
(oblinita),and mimicry of special objects, such as a spider’s nest 
(vulpina), or of some other specially defended larva (radcliffet 
mimics Datana or luteicoma, which probably mimics Notolo- 
phus (Orgyia) ). The type of larval structure is as usual in the 
higher Noctuids (Bombycoide), but there is also a considerable 
tendency to the development of secondary hairs (morula, dac- 
tylina, etc.). In one group there is a degeneration to the type 
of tubercle with single seta (hamamelis+), but true warts are 
present in the early stages. 
It will be seen that all the Acronyctz are excluded from the 
Lymantriide by the arrangement of the warts on the thorax; 
from the Notodontide by the presence of true warts, though 
sometimes degenerate ;. from the Arctiide they are not distin- 
guished. But these families, Arctiidze and Noctuide integrade 
even in the venation of the moths. The character which has 
been used to distinguish them (degree of coalescence of vein 8 
with subcostal on hind wings) is an unsatisfactory and incon- 
stant one, and we need not feel surprise that it does not corres- 
pond entirely with the development of warts in the larv: 
Thus Mr. Butler’s position appears to have been ill-founded. 
The structure and pattern of coloration of the moths are, in 
this instance, a better guide to their affinities than the resem- 
blances of the larvee—that is to say, the superficial resemblances. 
The structural characters of the larve confirm the usual classi- 
nenviowse as I have uae shown. 
= * See, how ever, a note by Prof. R. Meldola in Weismann’s ‘‘ Studies in the Theory of 
Jescent,”’ p. 169. 
P TL a species is a beautiful instance of the outogenetie stages repeating the course 
of phylogeny. 
{Several minor points show that the Acronycte belong with the Noctuids. For ex- 
ample, (1) no Arctian has see ar hairs, whereas these occur in the related Noctuid 
genera (¢. g., Panthea furcilla) ; (2) Abdominal segment 8 is characteristically enlarged 
in most Acronyct, and often be sa tubercles i and iiin a square. These characters are 
common to many: »Noe ‘tuids, but are not found among the Arctians. 
