EARLY STAGES OF SEGMENTATION OF ECHINUS HYBRIDS. 491 
roundj deeply stained bodies^ which are not always easy to 
distinguish from large yolk-granules or sometimes from 
normal chromosomes. Every stage may be found between 
these stained dots and true vesicles, and in the telophase the 
normal chromosomes within the daughter-nuclei give rise to 
chromatic dots exactly similar in appearance. The fate of 
the vesicles depends upon their position on the spindle; 
those which lie among the chromosomes of the anaphase 
groups are carried with them to the poles and become 
included in the daughter-nuclei, while those which lie on 
the periphery of the spindle or just on its equator remain 
where they are, and are excluded from the nuclei. They 
may commonly be found along the line of the cell divi- 
sion in the 2-cell stage, already considerably shrunken 
and reduced to stained dots. The number eliminated as 
counted in late anaphase figures before the cell-division 
varies from none to about five, but in 2-cell stages a larger 
number of chromatic dots may sometimes be seen, and it is 
possible that they break up as they undergo degeneration. 
Figs. 10-14 illustrate anaphase and telophase stages of the 
first segmentation division. 
Before proceeding to describe the secoml segmentation 
mitoses we must refer to a curious phenomenon which is 
very common in the first division figures. In a very large 
proportion of anaphases of the first division, and not rarely 
before the chromosomes have become arranged in the equa- 
torial plate or even immediately after the dissolution of the 
nuclear membrane, it is seen that the centrospheres of the 
spindle have become divided, so that what at first appears as 
a bipolar or more commonly quadripolar spindle results. The 
extent of the division varies; the two halves of the sphere 
may be close together so that the spindle-fibres are not 
seriously deranged (figs. 10, 11), or they may, even in pro- 
phase, be widely separated, and a small secondary sheaf of 
spindle fibres may sometimes be seen between them (figs. L5, 
16). At first we took these spindles with divided poles to be 
multipolar spindles caused by the entrance of more than one 
VOL. 58, PART .3. NEW SERIES. 33 
