Packard.] 
110 
[May 7, 
IX. HINTS ON THE ORIGIN OF THE RHOPALOCERA. 
The arrangement of the different superfamilies and families of 
the Lepidoptera has always presented great difficulties owing to 
the homogeneity of the order, both in the larval and adult stages, 
and entomologists have consequently entertained a great diversity 
of opinions, both as to the systematic position of the leading 
groups, and their natural limits. In studying the adult stage of 
the moths we have found that besides the venation, much depen- 
dence may be placed on the head-characters, and also on the shape 
of the thorax. In studying the adult larval stage, we are more 
or less baffled by the fact that the features which distinguish them 
are so generally adaptational. But a careful examination of the 
freshly hatched larvae, together with a knowledge of the entire life- 
history or ontogeny of the caterpillar will, if we are not mistaken, 
enable us to unravel some hitherto knotty problems. Some of 
these unsolved questions have been the probable origin of theBom- 
byces and allied groups of the Sphingidae and cognate families, 
but more especially that of the different families of the butterflies. 
There is indeed, as generally acknowledged, a slightly ascending 
series in the Lepidoptera ; the crepuscular forms naturally occupy- 
ing a space between the Bombyces and the Rhopalocera ; but be- 
yond this there is great uncertainty, and it is difficult to say which 
are the higher, or the more specialized, and whether all the fami- 
lies have arisen one after the other in serial order. 
I have found that the grouping of the imagines is comparatively 
easy if we take into account the relative shape and size of the 
ch 7 peus and epicranium, and the general shape of the thorax, but 
for obtaining some light on the origin of the different families, the 
lower or ancestral groups from which they have sprung, experience 
has taught us that we must study the freshly-hatched larvae. 
By a study of the ontogeny of quite a number of Bombyces of 
different families, we have, as already stated, satisfied ourselves that 
the Notodontidae represent the most archaic forms, and that they 
probably originated from forms more like the JNoctuidae than any 
other group. The Noctuidae, Phalaenidae (Geometridae) may have 
originated as more or less parallel groups from the Pyralidae, and 
the latter may have evolved from the Tineina, which are certainly, 
as all will agree, the most primitive and archaic group of Lepidop- 
tera. The Pterophoridae, so far from being the lowest Lepidop- 
