Marshall—On the Anatomy of the Dragonfly. 775 
instance where this appearance was noted the adjacent walls 
of the apparent nuclei touched. Often in surface view a sup¬ 
posedly binucleate cell could he seen by careful focussing to 
give a view in which these two apparent nuclei were connected. 
In many sections nuclei were seen of a greatly bent horse-shoe 
shape, the two prongs of the shoe being short and projecting to¬ 
wards the outer wall of the reservoir. The nuclei always lie 
in the large projections but may extend out into the other part 
of the wall. It appears that each cell really has but a single 
nucleus but that it is generally very much bent. 
While these two enlargements at the end of the ducts have 
been called the reservoirs of the salivary glands, and, from 
their position and connection, would appear to act as such, 
their structure would seem to give to them more than a passive 
function. In most of the specimens examined the cells showed, 
between the nucleus and the outer wall of the reservoir, a dis¬ 
tinct striation in the cytoplasm giving to them the appearance 
of active secretory cells. From the whole structure one would 
suppose that the secretion was expelled from the outer wall or 
that, if it found an exit into the reservoir there were special 
ducts to pass the secretion through the thick cuticula, lining 
the inner wall. Ho such ducts were observed, the only struc¬ 
tures that could in any way be taken for a collecting tube or 
passageway were small dark spots; these were rather regular 
in their appearance just where the inner comes nearest the 
outer wall (fig. 37, s). These were also seen as dark strands 
running from one cell to another parallel to the outer surface 
of the reservoir but could never be followed for any great dis¬ 
tance. A study of this so called reservoir of different species 
of dragonflies during periods of rest and activity might solve 
this problem. 
From the more rounded end of each reservoir a. small duct 
leads out, the two shortly uniting (fig. 35). The wall of this 
common duct has for some distance the peculiar covering of 
cells from the reservoir that was observed on the long ducts 
just before their end. Before opening externally the structure 
of the wall changes to a regular hypodermal layer with a thick 
