M. Koidzumi 
259 
Division takes place when the sphere and the nucleus are brought together, 
the sphere, the thread and the nipple almost simultaneously dividing into 
two. The axial column of the nipple splits longitudinally into two halves, 
the split beginning at the posterior end and extending forwards. The sur¬ 
rounding layer also divides, each half being accompanied by one of the 
daughter axes. The ball at the top of the column is not divided completely 
at this stage, but is constricted first into a dumb-bell, and remains connecting 
the top of the daughter axes until just prior to the completion of the whole 
process. When the little sphere at the end of the thread divides, the two 
daughter spheres are situated on the surface of the nucleus, each connected 
with the base of the daughter nipple by means of a daughter thread, formed 
apparently by a splitting of the original single thread. The most remarkable 
feature observable at this stage is a thick and distinct strand connecting the 
daughter spheres, on the surface of the nucleus (PI. XI, fig. 19). It is identical in 
its structure with the strand seen in T richonympha , though it is, in Pseudotri- 
chonympha, more conspicuous and distinctly fibrous. The above processes 
must take place very quickly, since I have not yet seen a specimen with the 
nipple already divided but the strand not yet found on the nucleus. Owing 
to its fineness and its faint stainability, the thread may frequently be in¬ 
visible, but the strand never escapes the observer’s eye. In this respect, 
Hartmann’s description appears somewhat singular to me. He described some 
individuals with divided nipples, but in none of these forms does he seem to 
have noticed either the strand or the thread connecting the nucleus with the 
nipple. It is certain, however, that only a few forms in an early stage of 
division were observed by Hartmann, and he did not follow the subsequent 
processes. Thus it is not clear whether a strand makes its appearance in the 
Brazilian species later than in our species, or whether it does not appear at all. 
The strand is now gradually elongated. The nuclear membrane persists 
throughout the whole division process, and the spheres are always attached 
firmly to it. Consequently elongation of the strand is accompanied by an 
alteration in the shape of the nucleus. The strand lies transversely and the 
nucleus is disposed, in the majority of cases, along the whole of its length, and 
appears as a segment of a circle (PI. XI, fig. 20). Hand in hand with the altera¬ 
tion of the shape of the nucleus, changes occur also in the chromatic network. 
Chromosomes are formed, by the chromatic substance forming an irregular 
network or a spongy mass, and coming gradually to assume the form of more 
or less distinct threads with nodular thickenings. These chromatic threads 
become gradually distinct and uniform in thickness, coiled in the nucleus, 
appearing like the spireme in metazoan cells. Finally they resolve themselves 
into a large number of chromosomes, arranged along the strand and fixed 
to the spheres at both ends. They then split transversely into pairs of daughter 
chromosomes. I think it probable that the chromosomes do not split simul¬ 
taneously, but are divided one by one. They are of nearly uniform thickness, 
but their surface is rough. The daughter chromosomes fixed to the polar 
17—2 
