71 
iods are already highly developed, indeed, it is for their use in this 
group that, so far as we can guess anything about it, they first 
originated. Of course, they really originated in some lost ancestors of 
this group, but it is hardly necessary to be continually stating such a 
well-known consideration, so for the practical purpose of tracing the 
history of these rods, we must begin with them here, obviously the 
basal point of the rod-possessing Lepidoptera. 
. regard to the Hepialids, we have a curious fact as to the 
spiracles that may or may not bear on our present subject, but has a 
certain interest in relation to, if not the Blues, at least the Hair- 
streaks. 
ihis fact is that in the Hepialid pupa the spiracles of 7th and 8th 
abdominal segments are obsolete or closed, but both are well developed 
in the male imago, and the 7th is so in the female, but in the female 
imago that of the 8th abdominal segment is quite obsolete and has 
left no trace. 
"With these and some other exceptions, it is the rule in Lepidoptera 
for the spiracle of the 8th abdominal segment to be wanting in the 
imago and closed in the pupa in both sexes, a circumstance that falls 
in line with many others in inducing us to regard the pupa as not as 
tiuly a separate instar as the egg or the larva, but as being in reality 
an immature imago. So far as I know, the Lycasnids follow this 
general rule as to the pupal and imaginal spiracles, except in ths case of 
the Theclas. In these the imagines follow the rule of there being seven 
pairs of abdominal spiracles, the 7th segment having them and the 8th 
being without. But in their pupae we find the spiracle of the 7th is closed, 
just as the 8th is, but is recovered in the imago, though the 8th is 
completely lost. 
In Eriocephala ( purpurella , etc.), we find that the 8th abdominal 
segment in the imago has a spiracle, both in the male and in the 
female, so that here the position is even more primitive than in the 
Ilepialids, where only the $ preserves this spiracle. Notwithstanding 
this, the elaboration of the last three segments is in other respects so 
gieat, that there must necessarily have been a very long evolution from 
such an ancestry as Mtcropteryx, and the structure does not give us 
that light that we hoped to obtain from so primitive a form. 
We have here the outer rods and 10th abdominal segment forming 
a remarkable instrument for piercing, and it seems impossible to resist 
the conclusion that the outer rods belong to the 10th abdominal 
segment. Nor have I met with any species showing structures that 
throw any serious doubt on this conclusion. 
The inner rods are, in Eriocephala, attached to the ventral plate of 
the 8th segment, the segment that carries the pairing opening, and I 
cannot decide what precisely represents the 9th segment. The other 
terebrant lepidoptera ( Incurvaria , Lampronia, Adelids, etc.) fail to give 
me much assistance, chiefly, however, because they are very difficult to 
dissect satisfactorily, but in some species there are minute plates on 
the membranous portion of the ovipositor between segments 8 and 10 
that can hardly be anything than representative of the plates of the 
missing segment, which must, therefore, be the 9th. If this is so the 
inner rods belong, as Peytoureau says, to the 8th segment, though 
xxii.-xxiii. 
