118 GERM-CELL CYCLE IN ANIMALS 
attempts seem to justify the conclusion that the 
division is equal or approximately equal, that is, 
each group contains about thirty-two germ cells. 
These groups acquire a covering of mesoderm cells, 
are carried by the somatic tissues to a position 
near the dorsal surface on either side of the body in 
the last two abdominal segments, and thus become 
germ glands situated in their definite positions. 
Some time before the larval stage is reached, the 
sex of the embryo can be determined by the shape 
of the germ glands; those of the male become dumb- 
bell shape, whereas the female organs retain the earlier 
pear shape and begin to acquire terminal filaments. 
It is interesting to note that much time and 
effort have been wasted by those who have attempted 
to influence the sex of caterpillars by over-feeding 
or starving. Kellogg (1907), for example, ‘“‘dis- 
covered,” after an unsuccessful attempt to change the 
sex of silk worms by this means, that these cater- 
pillars already possess germ glands which are dif- 
ferentiated as male or female. If he, and others 
who have undertaken similar experiments, had 
examined the literature on the origin of the germ 
cells in insects, they would have found that as long 
ago as 1815, Herold published results of investiga- 
tions on Papilio brassica and other species of Lrpt- 
DOPTERA which proved that the sex of the larva is 
already determined before it hatches from the egg. 
A similar condition was reported in Bombyx pini by 
Suckow (1828), in Zeuzera esculi by Bessels (1867), 
and in Pieris brassica by Brandt (1878). 
