1914} A Structural Study of Caterpillars. 119 
e probably represents normal g, while d would be a transverse 
fibre. The odd fibre e’ may be the last trace of the proleg. 
The posterior nerve, with the disappearance of the proleg, 
runs mainly to the skin. 
The last segment is specialized to such an extent that an 
embryological study would be necessary to straighten it out. 
It doubtless is a fusion of at least two. Its one large nerve 
runs mainly to the proleg and on its way serves as a stalk 
for the nerves of the preceding joint. 
COMPARATIVE ANATOMY. 
Including this paper I can find only five widely separated 
families represented by dissections of the muscular system, 
namely: 
Cossip#; Cossus cossus by Lyonet. 
NOTODONTID2; Pygera bucephala by Lubbock. 
LASIOCAMPID; Malacosoma americana and disstria in this 
paper. 
NocTuDID#; Two species, probably of Noctua and Nephe- 
lodes, both trifidee and a Pheocyma (quadrifide) in this paper. 
+  SPHINGIDA; Delilephila lineata, briefly by Berlese; and 
Sphecodina abboti in this paper. 
These few species indicate, however, that the characters 
of the muscles are likely to be as well marked as of any other 
part; the following points stand out most strikingly: 
In the Sphingide the muscles, especially the large ones, 
tend to be broken up secondarily into a considerable number 
of fibres. This shows strikingly in the recti. Q and R show 
.a variety of lengths and tend to run to at least two secondary 
annulets; x and p cross the middle line and interlace with each 
other. Well marked resemblances to Cossus appear in the 
broad convergent and overlapping e and ff, in the filling of the 
body cavity with tracheze, and the slight differentiation of 4. 
The thoracic muscles are particularly massive; the longitudinal 
connectives are fused unusually far in the thorax, and com- 
pletely in the abdomen, throwing the apparent origin of the 
sympathetic fibres back to the ganglion of the next segment. 
The Cosside show their primitive character in the well 
separated ganglia in the seventh segment, the slight differentia- 
tion of 4, the fact that x is nearly longitudinal and hardly 
distinct from t, even in the segments with prolegs, and the well 
developed ninth segment of the abdomen. 
