ON SPERMATOGENESIS IN THE RAT. 
359 
contains the accessory corpuscle and some fatty granules 
remains bulged (figs. 18 and 23). The filament which joins 
the nucleus to the cilium is more plainly seen in the narrow 
part of the cell. This lengthening of the spermatozoon is 
accompanied by a corresponding movement of the heads 
downwards along the protoplasm of the supporting cell, until 
when the spermatozoa have attained their full length their 
heads reach the outer layer, in the neighbourhood of the 
nucleus, where they remain until the spermatozoa are cast 
off into the lumen. A group at this stage is represented 
by figure 19. The middle piece of the spermatozoon is now 
clearly visible passing through the protoplasm, which is col- 
lected chiefly at its upper end near the junction with the tail. 
The middle piece is formed out of the protoplasm of the cell, 
but not, as might be supposed, from the whole of the protoplasm, 
for a residual portion separates off from the spermatozoon, 
when its development is nearly completed. This residual part 
of the protoplasm, which contains the accessory corpuscle and 
one or two clusters of small fatty granules, gradually accumu- 
lates in the form of a globule, which separates from the body 
of the spermatozoon. At first the globule remains attached to 
the upper part of the body by a short pedicle (a group of 
spermatozoa at this stage is represented by fig. 20, and columns 
of spermatozoa in situ with the globules attached in fig. 15), 
but before long it breaks away entirely from the spermatozoon. 
In sections stained with haematoxylin small chromatic gra- 
nules make their appearance in the columns of spermatozoa 
(fig. 9, x), and when the spermatozoa have passed into the 
lumen these granules, each of which is contained in a small 
amount of protoplasmic material, are found detached from the 
columns, and occupying the interval between the heads of the 
spermatozoa and the cells of the third layer. These bodies 
have been previously described as the seminal granules ; they 
are, in fact, the globules which, as we have just seen, separate 
from the spermatozoa at a late stage of their development, and 
the chromatic granules seem to be formed in part by the clusters 
of fatty granules seen in osmic acid preparations (figs. 10 and 
VOL. XXV. NEW SER. 
A A 
