210 BOTANICAL GAZETTE [SEPTEMBER 
longitudinally to the ovule and therefore at right angles to the 
equatorial plate, it is difficult to make an accurate count of the 
chromosomes at a later stage of the mitosis. It is certain, however, 
that we are here dealing with the haploid number, probably 12. 
In vegetative divisions 20 or more chromosomes can easily be 
counted, though in this case also it is not easy to determine the 
number with certainty. The chromosomes when drawn into the 
equatorial plate are short and thick, almost globular, strongly 
contrasting in this respect with the long, crowded chromosomes 
of the vegetative division. 
In none of my material, though careful examination was made, 
could there be found any difference in the separating chromosomes 
(fig. 3). Those going to the lower pole are fully as large and give 
the same staining reaction as those going to the upper pole. But 
on arriving at the poles of the spindle, the two groups behave very 
differently. Those at the upper pole unite into a normal nucleus, 
while those reaching the lower pole fuse together into an irregular 
lump, without spongioplasm or distinct nuclear membrane. Fre- 
quently chromosomes or chromosome fragments fail to reach the 
principal mass, and remain scattered along the spindle or in the 
cytoplasm outside. When Flemming’s triple stain is used, these 
fragments, as well as the large chromatic lump, take only the 
safranin and appear semitransparent, while the chromatin of the 
upper nucleus, taking the gentian violet, appears dark and opaque. 
No wall is run in at the close of this division, but a distinct cell 
plate is formed by the thickening up of the spindle fibers (fig. 4). 
Division follows in each of the daughter nuclei resulting from 
the heterotypic mitosis. My material, however, does not furnish 
any examples of the prophase of this division, nor any information 
with respect to the chromosomes. The telophase is represented in 
figs. 5 and 6. In both figures the remains of the first spindle and its 
cell plate are still distinguishable (on the left side of fig. 6). A cell 
plate is formed also on the second spindles, but these and the earlier 
cell plate are transient structures and disappear shortly. 
The division of the upper nucleus results in this case also in the 
formation of an upper healthy nucleus and a lower irregular lump 
of chromatin. By the division of the lower of the daughter 
