THE EYE AND OPTIC TRACT OF INSECTS. 
223 
The terminal anastomosis of Agrion may be conveniently 
divided into four regions. First, the region (1) lying nearest 
to the peri-opticon in which the nerve-cells are numerous, and 
the fibrils leaving the peri-opticon form a complicated plexus, 
the region (2), next to this, in which the fibrils have collected 
into bundles separated by spaces occupied by very thin- 
walled tracheae in which there are no spiral markings, and lymph 
spaces; next, the region (3), in which the fibrils form a final 
plexus, and in which there are again a considerable number of 
nerve-cells, and, lastly, the region (4), in which the fibrils are 
again collected into bundles, separated by spaces containing 
tracheae, which perforate the basement membrane to supply 
the retinulae. 
In Noctua, Sphinx, and Acherontia, the three genera of 
Lepidoptera I have examined, the peri-opticon is composed 
of the usual cylindrical elements, but they are much longer, 
thinner, and more tightly packed than usual, so that the whole 
ganglion has a much more compact and spherical appearance 
than it has in any of the genera we have hitherto considered. 
The peri-opticon is closely connected with the epi-opticon 
by a thick nerve tract in which the nerve-fibrils completely 
decussate, but neither in this region nor in the epi-opticon nor 
in the peri-opticon is there any trace of tracheal vessels per- 
forating the tissues. 
The terminal anastomosis of the Lepidoptera is most extraor- 
dinarily complex, and the four regions described above in Agrion 
can be readily distinguished (conf. Leydig, Taf. x, fig. 2). In 
region 2, the large thin- walled, but still spirally-marked tracheae 
may be readily seen branching between the nerve-fibril bundles. 
The terminal anastomosis of the Lepidoptera is usually very 
deeply pigmented. In the bee and the wasp, the only two 
members of the order Hymenoptera I have examined, the 
peri-opticon is very similar to that of the Lepidoptera, the 
elements being long, delicate, and very close to one another. 
The terminal anastomosis is not so complicated, nor is it 
usually so densely pigmented. No spirally-marked trachea} pene- 
trate the optic tract at any part of its course in Hymenoptera. 
