103 
[Packard. 
and ninth abdominal segments. As the thoracic spines are branched, 
as well as the median ones at the end of the body, it does not seem 
so probable that the latter have arisen from double spines, as sug- 
gested by W. Muller, and yet on the whole his view seems to us 
correct. 
In the Attaci there is some variation in the spines on the eighth 
abdominal segment. In Platysamia cecropia the larva of the first 
stage lias on this segment a single spinulated rounded tubercle ; in 
Actias luna this tubercle is also single , 1 but in Telea polyphemus and 
Callosamia promethea in stage i there is a double twinned tubercle , 
ivliich in stage n becomes single ! This fact indicates that the median 
single spine may be the result of the coalescence of what were orig- 
inally two spiniferous tubercles . 2 
We will now turn to the Ceratocampidse. We have drawings of 
the different larval stages of Eacles imperialism Citheroniaregalis and 
JS phingicampa bicolor as well as some of the early stages of Ani- 
sotaand Dryocampa, made under our direction by Mr. Bridgham . 3 
Beginning with perhaps the lower, more generalized forms Dryo- 
campa rubicunda , Anisota stigma and A. senatorial in both the 
earliest and latest stages there are two remote spines on the eighth 
abdominal segment, with a single one on the ninth. This latter spine, 
however, in the first and second stages of A. senatorial arises (if Mr. 
Bridgham’s drawings are correct, for I have at present no specimens 
by which to correct them) from a larger wart than those adjoining, 
which gives rise to two bristles. It thus appears that this spine is 
double in its origin. But this conclusion needs further confirmation. 
In the third and later stage, the two-bristled wart is represented 
by a solid, sharp spine, both in Anisota and in Dryocampa, while 
1 Since this paper was presented to the Society I have bred both P. cecropia and A. 
luna. In P. cecropia the tubercle in question is about twice as large round as the others 
on the same segment, and is evidently the result of the concrescence in the embryo stage 
of two tubercles; it bears eight to ten bristles or nearly twice as many as the homolo- 
gous tubercles on the other segments. That of Samia cynthia was originally double 
being broad and bearing four bristles on each side. That of A. luna. is more markedly 
double than that of Platysamia cecropia , having five setae on each side of the summit. 
2 VVe have noticed that in stage I of Smerintlius exccecatus the caudal horn is distinctly 
forked at the tip, and that in stage ill it ends in two tubercles. In Sphinx kalmice in 
the second or third stage the horn ends in three or four tubercles but in stage iv? 
there are no definite traces of a fork. In a Sphinx found on the larch the smooth horn 
ends in two fine setae. In a lot of freshly hatched Sphinx larvae of unknown species 
the horn is distinctly forked. 
3 Since this paper was sent to the printer, we have, during the past summer', observed 
and reared all these larvae, and possess additional pi oof of the duplex origin of the 
caudal horn. 
