30 MOSQUITOES OF NORTH AMERICA 
fig. 1), especially in the male. As mentioned above, the shaft of the antenna 
(pl. I, fig. 1, C”) is inserted in the center of a plate which is striped with radial 
thickenings of the cuticula (fig. 1, P; fig. 2, P) and which forms the floor of the 
cup-shaped depression upon the distal surface of the segment. Through an 
opening in its center the tracheal trunk and the nerves (fig. 1, V”, T’) penetrate 
into the shaft of the antenna. Around the base of the shaft the plate is 0.01 mm. 
in thickness, however, towards the periphery it becomes somewhat thinner, then 
again thicker. At the same time it bends forward and joins the chitin covering 
the sides of the cup-shaped depression. This plate therefore, when seen from in 
front, has the form of a radially striped saucer with an opening in the center. 
From the periphery of this plate, in its entire circumference, there arise in a 
lateral direction 70 to 90 fine chitinous processes, 0.03-0.05 mm. in length, 
which gradually bend forward (fig. 1, A, 2, 4) and which serve as points of at- 
tachment to the nervous end-organs in the manner to be described later. The 
processes do not take their origin from the surface of the plate but from deep 
within its substance. This is proved by the fact that in a section through this 
part of the plate from front to back the transverse sections of the processes can 
be recognized as small, more strongly light difracting circles (fig. 3, 4) within 
the surrounding mass of the plate substance. The plate is composed of cuticular 
substance and its posterior surface is covered by a layer of hypodermal cells 
(fig. 1, H'). The plate is affected by different coloring substances in entirely 
the same manner as the connecting membrane between the antennal segments ; 
for this reason and others to be mentioned below, it is not to be looked upon as 
a separate segment of the antenna but simply as connecting membrane. 
The outer surface of the segment, as well as the sides of the cup-shaped de- 
pression up to the margin of the plate, which also belong to the surface, is formed 
by a layer of chitin, 0.005—0.008 mm. in thickness (fig. 1, C’). At the anterior 
margin of the cup its hollow is somewhat narrowed by the thickening of the 
chitin (fig. 1). At the base also, which serves as point of insertion for the 
muscles moving the antenna, the chitin is somewhat thickened. Under the 
chitinous layer there lies the hypodermal layer which in the imago generally 
consists of very small cells (fig. 1, H). 
“ Separated from the chitinous covering by a narrow space there is a layer of 
ganglion cells, 0.02—0.04 mm. in thickness (fig. 1, @),* consisting of nearly 
round, indistinctly defined ganglion cells of about 0.005 mm., which possess 
comparatively large nuclei. In preparations by maceration one sees that from 
the cells fine protoplasmic processes proceed (fig. 5, G). Owing to these fine 
threads, which are probably connected with each other, this layer has the ap- 
pearance, under low magnification, of a nearly homogeneous mass densely filled 
with nuclei. 
“ Still farther inward, between the ganglion cells and the end-organs, there 
lies a layer, 0.005-0.007 mm. in thickness, of fibres running in two directions 
(fig. 1, F; 4, F). One set of these connects the ganglion cells, with the end 
organs, the others cross these at right angles from behind forward; these are 
fibres originating from the large antennal nerve and which connect with the 
ganglion cells. On the posterior portion of the ganglion layer the place can be 
detected where this nerve fiber passes through the cell layer (fig. 1, N’). It is 
noteworthy in this connection that the nerve fibres, which supply the larger part 
of the cell layer, lie between these and the end-organs instead of on the outer 
side of the layer, as in the posterior section. The smaller portion of the ganglion 
cells lying within the posterior half of the organ (fig. 1, G’) connect mostly with 
fibres emanating directly from the nerve. 
*“ With reference to the cells my drawings are diagrammatic. I have not attempted to 
draw in the fine processes.” 
