Marshall and Severin — Anatomy of Ranantra fusca. 493 
enlargement similar to that figured by Dufour (2) and by 
Locy (5), but not nearly so great. 
A closer examination of one of these tubes shows that it can¬ 
not be a non-glandular portion of the silivary glands but is either 
an active part of this gland or a, separate one, the use of which 
we could not determine. There are! three distinct 1 parts of the 
tube, the enlargement near th ganglion and the thinner parts 
lying in front of and back of this. The distal portion which ex¬ 
tends through most of the metathorax and the first two abdominal 
segments is a narrow tube having the lumen about as thick as the 
two walls. PL XXXIV, Fig. 10 D 1 .) The wall is composed of a 
single layer of what are apparently secretory cells resembling in 
general the structure of a, Malpighian tube. The proximal por¬ 
tion, extending through the meso- and pro-thoacic segments is the 
narrowest part and closely resembles the ducts leading from the 
salivary glands. It extends forward into the head, but its out¬ 
let we were unable to discover. 
The enlarged middle portion is peculiar in its structure and 
differs from either of the other parts. A view of a mounted 
specimen shows, us that the distal portion gradually passes into 
it with very little change in the general appearance except an in¬ 
crease in size. There are, however, a number of longitudinal 
strands passing partially through it. (PL XXXIV,Fig. 10 Str.) 
Towards its proximal end these strands in part disappear, the 
cells become smaller and more closely packed together, and this 
entire part is stained very much darker than, any other portion 
of the tube. A longitudinal section of this median portion (PL 
XXXV, Fig. 11) shows the wall to be continuous throughout. 
The strands pass forward into the darkly stained part where 
they apparently become closely packed together and lost. The 
cells and the nuclei become smaller and a mass of cells is formed 
which fills the lumen and even extends for a short distance into 
the tube. There is certainly no. structure similar to this in the 
salivary glands, and whatever the function of this tube may be it 
does not act as a reservoir for the salivary glands, but is more 
likely a part of the salivary gland or a special gland. 
Respiralory System. —(Pl. XXXV, Fig. 12.) One of the most 
noticeable features of Ranantra is the presence at the posterior 
end of the two long caudal filaments which together form the so 
called “respiratory tube.” As already explained, when the in¬ 
sect rises to the surface the posterior end is much higher than the 
