136 
Charles E. Allen 
membrane (Figs. 12 — 15). A proximity to the nucleus in certain cases 
of structures very similar in appearance to tliese has been taten as evidence 
of their origin from the nuclear substances; but in the present instance, 
as we have seen, this proximity follows naturally from the occurrences 
of the preceding mitosis, and affords no reason for assimiing a nuclear 
origin of the kinetosomes. 
The division into two groups seems in general to result from a Separa- 
tion of the kinetosomes already present rather than from an increase in 
their number. It is true that, supposing a division of the individual bodies 
to occur, conclusive evidence of the fact would be difficult to obtain because 
of the great differences among them in size and form. Their number, 
too, is variable, seeming to differ, for instance, even as between the two 
groups within a single cell. In general, the number of the kinetosomes 
is smaller in the later androgonial generations than in the earlier ones 
in which they appear (compare the smaller, latcr-formed cells of Figs. 20, 
21, 23 and 24, PI. VI, and Figs. 38 — 40, PI. VII, with the larger cells of 
Figs. 11 and 16, PI. VI), an observation which Supports the notion of a 
Segregation of entire bodies preparatory to each mitosis. However, my 
preparations do not show such differences in number between earlier and 
later generations as would exclude all possibility of a division of the indi- 
vidual bodies; and if, as appears to be the case, the groups of kinetosomes 
result from a breaking up of the polar plates of the earlier generations, 
an occasional further fragmentation of some of the smaller bodies would 
not be surprising. Sometimes the number of kinetosomes at either 
spindle pole is the same or about the same as the chromosome number 
(PI. VIII, Fig. 65), but such coincidences are probably not significant. 
As the daughter groups move apart (PI. VI, Figs. 12 — 15), some of the 
kinetosomes often lag behind their fellows, and even in the equatorial 
plate stage (PI. VII, Figs. 38, 39) afew of them may be scattered irregularly 
through the cell. The majority, however, are always clearly referable 
to one or the other group. 
When the diverging groups have reached their final positions (Figs. 
16 — 18), some of the fibers of the future spindle appear. Fibers have 
not been found in any case while the kinetosomes are moving apart, as 
they were at corresponding stages in the Separation of the polar plates. 
Negative evidence, to be sure, is not conclusive on this point, since fibers 
running to the nuclear membrane in cells like tliose shown in Figures 12 — 15 
might well be difficult of detection because of the close grouping of the 
kinetosomes. It is true, however, that in the cells which contain kineto- 
somes the amount of fibrous material, as well as of polar kinoplasm, is 
