24 The Developmental Characters of
'Synonymic List,' and by Newman in his 'British Moths,'
appears to be most unnatural.
It has often been asserted by Lepidopterists that no true
systematic arrangement can be founded on larval characters
alone, and this remark is to some extent correct, if we limit
our comparisons to the adult caterpillars only, so that objec-
tions might perhaps be legitimately raised from this point of
view to the suggested restoration of the older arrangement.
But the case is quite different if, instead of comparing the
adult larvae, we study their comparative ontogeny. By this
means, as explained in the previous part of this paper, we
may gain an insight into the true relationships of the species.
The first application of this principle to Lepidopterous larvae
was made by Professor Weismann, of Freiburg, in his
'Studien zur Descendenz-Theorie' (vol. ii., 1876), in which
work the author has studied the markings of the larva of the
Sphingidae in great detail from this point of view. The whole
subject is as yet almost in its infancy, and one of my chief
objects in bringing this paper before your notice is to
endeavour to convince you of the great necessity for paying
attention to the earlier stages of caterpillars. No group of
butterflies or moths has as yet been completely worked out
in the manner in which Weismann has investigated the
Sphingidae, and till this is done for the whole order we shall
never arrive at a truly natural classification. By making
careful drawings of caterpillars at every stage of growth—
using the microscope in the case of the youngest stages—all
who are given to larva breeding may contribute substantially
to the more scientific portions of our studies.2
I shall now adduce evidence to show that the larvae of
most of the Bombyciform Nocture are geometriform when
young; the Geometriform Noctua are, as will be remem-
bered, generally semi-loopers throughout life.
Among the earliest observations in this country is a note
by Dr. Buchanan White (Ent. Mo. Mag., vol. v., 1869,
2 The genus Acronycta specially claims to be thus worked out. See A.
G. Butler in Trans. Ent. Soc, 1879, p. 313.