20 
abdominal segments of those pup® which have the power of moving 
from place to place are also separated by movable incisions, the earliest 
pup* (of the more quiescent form) would also have more or less mov¬ 
able abdominal segments. ’ We may grant then that the most ancient 
form of lepidopterous pupa did have more or less movable segments, 
and that, as development has proceeded in certain species, the pup* 
of such have slowly lost the power of movement, first one incision and 
then another becoming fixed, retaining, however, such movable inci¬ 
sions as natural selection indicated as being of service. In some 
pup* the power of movement is absolutely lost, all the segments of 
the abdomen are firmly welded together, and such a solid pupa 
undoubtedly represents the most recently developed and most specialised 
evolutionary product the pupa can present in this direction. It is on 
these lines that Dr. Chapman has largely worked, and on which, 
coupled with other structural peculiarities, he formulated the revolu¬ 
tionary and sweeping changes in classification, which more recent 
studies in other directions have proved to be well founded. 
Characters from the perfect insect have been almost exclusively 
used by museum-workers and closet naturalists. Living things are, 
in most entomological museum-rooms at least, quite out of place, 
hence caterpillars and pup* are conspicuous by their absence. 
Practically, too, until the last quarter of a century the important 
character of the egg, larval and pupal stages, were not recognised 
or understood, hence entomologists had to fall back on the imaginal 
characters. Palpi, maxill*, antenn*, legs, form and shape of wings, 
have all been used for the purposes of classification, but the latest 
and perhaps most important character that has been largely used is that 
of neuration. 
So far, however, as we have got, it is unsatisfactory to find that 
neuration has not yielded any results that can be placed on the same 
level as Dr. Chapman’s work on pup*, or Chapman’s and Dyar’s 
work on larv* ; and this seems to be, not so much on account of its 
not being quite as useful a character for the purpose as those they 
have chosen, as on account of the fact that we have not yet found the 
man who can read its riddles so satisfactorily. So far, Comstock and 
Hampson have done the best work therein, but there is plenty of room 
yet for either to reach a much higher level than he has yet done. 
It appears very doubtful whether anyone has yet properly discovered 
the true phylogenetic meanings which neuration conveys, or whether 
anyone has examined sufficient material to certainly determine what 
neurational modifications are really of phylogenetic value, and what 
are simply variable factors, and of no classificatory use. In these 
studies the method of numbering the nervures has proved a pitfall, 
and when we find the lower branch of the subcostal nervure of Eucliloe 
carclamines marked with the same number as the upper branch of the 
median of Papilio machaon, we feel there is something radically wrong. 
This leads me to say that the most important character of neuration, 
phylogenetically, is not so much the number of nervures around the 
discoidal cell as the method of formation of the cell itself, for if this 
were considered it would be impossible to suppose that the branch of 
the subcostal in Euchloe is homologous in any way with a branch of 
the median in Papilio. 
