M. Koidzumi 
287 
nothing to say with certainty. At the time when they first become visible 
in my preparations, no centrioles can be seen at the points where the flagellar 
cords join the nuclear membranes. 
I have found no other processes of multiplication than those just described. 
The stages of division of Pyrsonympha (incorrectly called Dinenympha gracilis) 
described and figured by Zulueta (1915) agree closely, so far as they go, with 
those which I have observed. On the other hand, the remarkable cycle of 
development described by Comes (1910 a, 1912) in similar forms is very 
different from anything that I have seen. This author describes “males” 
and “females,” “conjugation,” “multiple fission,” and a peculiar develop¬ 
ment in the salivary glands of the termites. I agree with Grassi and Zulueta 
in believing that no such stages exist, and that Comes’ interpretation of his 
findings is incorrect. 
As mentioned already a remarkably large number of individuals are found 
attached to the intestinal wall: and these are not only young and small forms, 
Text-fig. D. A late stage of division of P. grandis, showing the flagellar cords and newly formed 
axial filaments. Drawn with camera lucida. (4 x T V.) 
* 
but large and mature ones as well. In studying these forms I often noticed 
gigantic piriform or fusiform individuals with a conspicuously jagged surface, 
distinguishable at a glance from the normal forms. On making a special 
search for these peculiar forms in a large number of termites, I was able to 
find not a small number of them, representing several types of structure. 
At a time when I had insufficient material, I was inclined to imagine that 
they represent stages in a particular mode of multiplication. The fact that 
a large number of young forms are found hanging from the wall in close 
association seemed to favour this supposition. But on studying more material, 
I soon realized that they represent successive stages in a process of physio¬ 
logical degeneration. 
In the early stages of degeneration the nucleus becomes detached from 
both the axial filament and the body wall; in some forms it is broken in two, 
and one half remains connected with the filament at its normal position, while 
the other is set free in the endoplasm (PI. XIV, fig. 62). The endoplasm itself 
