230 BOTANICAL GAZETTE . [OCTOBER 
longitudinal diameter, the spindle also appears very much flat- 
tened (fig. 2r). Were such a spindle sectioned it could easily — 
give the appearance of a multipolar structure. No such cases, 
however, were found. Ifthe spindle extended clear across the 
cell so that it touched the opposite walls it might give the 
appearance of the threads ending in the cell wall. Although 
the nature and origin of the spindles in figs. rg, 2z and 22 are 
exactly the same, there is a striking difference in their shape, 
and very suggestive of how the shape of the cell may influence 
the appearance of the karyokinetic figures. The same is evident _ 
from a comparison of figs. 6 and ro. Often the two poles of a 
spindle are not 180° apart. This is caused no doubt by the 
centrospheres not becoming exactly opposed (fig. zg). After 
the nuclear membrane has disappeared the V-shaped chromo- 
somes are gradually drawn down into the equatorial plane, with 
their heads toward the center, until they form quite a symmetri- 
cal figure ( figs. 22-26). The centrosome usually does not 
divide until after the formation of the mother star, but some- 
times the division may occur earlier ( fig. 23). The longitudi- 
nal splitting of the chromosomes takes place about or during the 
time of the formation of the mother star (figs. 27, 28). When 
the cell is very long and narrow there does not appear to be @ 
typical mother star formed ( fig. 24). Insuch cases the chro- 
mosomes do not appear to be drawn symmetrically into the 
equatorial plane. The chromosomes appeared quite homogene- 
ous throughout, nothing being visible having the appearance of 
chromatin granules. It was not possible to tell exactly how the 
chromosomes are arranged on the achromatic spindle threads, 
but the threads seemed to be in bundles running continuously 
from one pole to another, ending in the hyaline area of the cen- 
trosphere and having the chromosomes attached by their heads | 
( figs. 28-30). 
When the chromosomes have been brought into the equato- 
rial plane, and longitudinal splitting is complete, the daughter 
chromosomes are gradually pulled apart, and the central spindle 
begins to appear between them in the equatorial region. Some- 
