88 
&c. Although it is a little outside the scope of this paper, yet it is of 
interest to note that these earliest forms had great variability in 
respect of antenme. We find plumose, or at least strongly pectinated, 
antennae in some of the earlier Cossids; minute antennae in Hepialus ; 
very long ones in Adela. The latter is a good instance of a similar 
development separately attained ; certain Trichoptera have very 
similar antennae, and we find the connecting link in Micropteryx which 
has ordinary antennae, but at the same time possesses the pupal jaws of 
the Trichoptera and the ovipositing knives of Adela. 
Now it was somewhere here, amongst these forms with variable 
antennae, that the first trace of butterflies with their clubbed antennae 
appeared, and some little distance along the road there branched off, 
as a record thereof, the Sesiidae. These latter are unmistakeably 
“ Micros,” not, indeed, very far removed from Hepialus and Cossus, 
but they have fairly clubbed antennae. Some considerable distance 
further on we have Castnia , its larva still an internal feeder, and its 
pupa* still of a Cossus- like “Micro” character, but its imagines so 
very Hesperid-like that they have by some authorities been placed 
with the butterflies. There are other families that are possibly 
appendages of this line of evolution, but I know so little about them, and 
especially about their pupae, that it is prudent to say nothing further 
about them at present. Clubbed antennae are found in Sphinges and 
in Zygaena; in both cases they appear to have been acquired 
independently of the butterfly stirps, and of each other. This circum¬ 
stance seems to lend additional force to the idea that clubbed antennae 
are in some way specially useful to diurnal species. 
To sum up the points in this paper. My chief aim has been to call 
attention to the study of pupae, and especially those of butterflies from 
a broader and more general point of view, and to bring to bear on 
them the general principles of pupal evolution that were suggested by 
my study of the Heterocera. The special facts brought out concern¬ 
ing butterfly pupae are to be taken as largely preliminary and tenta¬ 
tive, but it is to be noticed that broadly, and even in some detail, the 
relationship of the different families to each other suggested, 
agrees with the ordinary classification. The greatest change of views 
which appears to be demanded is in relation to the position of the 
Lycaenids, which should no longer be regarded as in any way inter¬ 
mediate between the Papilionids and Nymplialids. Rather should the 
Lemoniidae and Lycaenidae be regarded as a branch which developed 
from the primeval butterfly (above the Hesperids) in one direction, 
whilst the Papilionids arose and branched to the Pierids and 
Nymphalids quite independently, f Another jioint is that the Pierid 
separated from the Papilionid at a very early stage of the evolution of 
the latter, and that the Nymphalid almost immediately thereafter 
separated from the Pierid, at a time, that is, when each yet retained (as 
many genera do still) both the 5th and 6th abdominal segments free, 
and before the Pierid had definitely acquired a single instead of a 
double nose-horn. 
* I am sorry that I have to depend on figures for my knowledge of these. 
f I do not wish this to he understood as denying that Papilio must 
nevertheless be taken as the nearest representative we have of the primary 
butterfly. 
