355 
summit, where it is continuous with the overlying corneal facet 
(ie..2, ps. c.). 
In Vespa during part of the pupal stage, the cone cells lie at 
the bottom of a considerable depression, and their outer ends 
are conspicuously pointed, as in Belostoma. Over the 
ommatidia, are faintly stained, rounded bodies which I formerly mistook 
for nuclei of the corneagen (fig. 3, p. s. c.). But further study shows 
they are chitenous bodies secreted by the cone cells 
and that they probably correspond to the pseudocone 
in Tabanus and Belostoma. 
Fig. 1. Semidiagrammatic view of an ommatidium of Belostoma. 
Fig. 2. Same, of Tabanus. 
Fig. 3. Same, of young pupa of Vespa: A—D. Cross sections of ommatidium 
and points whence they were taken. a—c. Same of Tabanus. E. Semidiagrammatic 
view of cone cells in an old Hymenopterous pupa. azx.n., axial nerve bundle. c.c., 
crystalline cone cells. c.g., incipient corneagen. g1—4, incipient ganglion cells. 2, nerve 
branch. o.p., optic pit. o.s., ommatidial spine. o.¢., cuticular tube, formed by circle of 
cells around the ommatidium. p.g., primary pigment cell. ps.c., pseudocone. r.A., rudi- 
mentary hairs on cone cells. ¢., trachea. 
The presence of these hair-like pseudocones over the cone cells of 
Belostoma, Tabanus and Vespa cannot but awake the sus- 
picion that the ommatidia are modified hair-bearing organs, and this 
suspicion is fully confirmed by the fact that, in the young pupae of 
Vespa, the first corneal cuticula is actually provided with hair-like 
spines unquestionably formed by the hardening of the outer ends of 
25* 
