1198 
Journal of Agricultural Research 
Vol. XXVIII, No. 12 
rect in several particulars. The “cloison m6diane” is restricted to the anterior 
segments of the trunk; it does not form a continuous membrane; it does some¬ 
times become attached to the muscular coat of the mid-intestine; it does not 
divide the fat body, since this is already divided mesially by a space or sinus 
dorsad to the mid-intestine; and it does not appear to be continuous with the two 
halves of the dorsal diaphragm, but represents delicate processes of connective 
tissue cells attached to the ventral surface of the heart. It is apparently absent 
in mature larvae and is probably a structure of minor importance. 
Structurally the dorsal diaphragm consists of two very delicate membranes, 
having the appearance of basement membranes and possibly being chitinous. 
These membranes are attached to the ventral wall of the heart in the mid-line. 
Between the two membranes are situated the diaphragm cells (fig. 5, A, DphCl). 
These are small cells, quadrangular in form, each having a spherical nucleus 
which occupies the greater portion of the cell. The diaphragm cells are arranged 
in single rows or strings which are bent in a sinuous form and frequently anasto¬ 
mose, forming a sort of lacework and giving the dorsal diaphragm in surface view 
a curious vermiculated pattern (PL 7, A). These cells are not found anterior to 
the middle of the fourth abdominal segment. The diaphragm cells are closely 
covered on their dorsal and ventral faces by the two membranes, which here have 
the appearance of limiting membranes secreted by the cells (PI. 7, B, DDph ). 
Elsewhere, in the open spaces between the diaphragm cells, the two membranes 
are generally evident, but are often found in close apposition, giving the appear¬ 
ance of a single membrane. 
The diaphragm cells have a general resemblance to the blood cells in size and 
staining reaction, and in the large size of the nucleus as compared with that of 
the cell body. For this reason Anglas ( 1) has assumed that the diaphragm cells 
give rise to blood cells (leucocytes). This may perhaps be possible, although the 
diaphragm cells are as a rule slightly larger than the blood cells. Moreover, the 
latter show such frequent mitoses that an independent origin for them need dot 
be assumed. 
Accompanying the dorsal diaphragm and forming an integral part of it are 
extremely delicate muscle fibers (fig. 5, A, DphCl), the so-called alary muscles, 
apparently unicellular and corresponding histologically to the fibrillae of other 
muscles. These fibers or fibrils are attached at more or less regular intervals to 
the ventral wall of the heart, the line of attachment coinciding with that of the 
membranes. The fibers corresponding to each chamber of the heart converge to 
their intersegmental points of attachment of the diaphragm on the body wall, 
thus giving the fibers the well known fan-like arrangement (PL 7, A). The rela¬ 
tion of the diaphragm muscles to the membranes of the diaphragm does not seem 
to be uniform. At some points the muscles appear to run between the two mem¬ 
branes, at others dorsad to them. Thus far the description of the structure of 
the dorsal diaphragm of the larva corresponds with the account given by Snod¬ 
grass (46) for the imago. In the latter, however, the dorsal diaphragm termi¬ 
nates in the 7th abdominal segment, since the segments remaining are highly 
modified. The histological elements are the same in both, except that this struc¬ 
ture in the imago is much more robust than in the larva, and can easily be seen 
on dissection, while in the larva it is so delicate that special methods are necessary 
to demonstrate it. In regard to one feature, a discrepancy exists between the 
two accounts. Snodgrass speaks of the diaphragm cells as situated upon the 
membrane, whereas in both cases they undoubtedly lie between two membranes. 
The dorsal diaphragm anterior to abdominal segment 4 is less well defined 
than in the segments posterior to this point, and does not form a continuous 
sheet, being entirely absent in the middle of the segments and restricted to tri¬ 
angular areas corresponding with the diaphragm muscles (Pl. 7, A). These 
