34 BOTANICAL GAZETTE [JULY 
above the other and are both quite large, each containing a 
large chromatin mass in the center (fig. 79). The lower one 
now divides into three daughter nuclei, all very much smaller 
than the parent nucleus, and with the chromatin scattered instead 
of being aggregated in a ball. These are all to be considered as 
antipodal cells, and persist in the little pouch until after fertili- 
zation. The upper nucleus continues to enlarge somewhat until 
about the time of fertilization, when it undergoes division fol- 
lowed by a cell-wall, as a result of which one daughter nucleus 
is inclosed in a cavity at the antipodal end of the embryo-sac 
(fig. 2r). This then must be considered as a fourth antipodal 
cell, since it is one of the four parts of the original lower 
nucleus. The other daughter nucleus is the polar nucleus, but 
becomes at once the mother cell of the endosperm, and imme- 
diately undergoes division until a parietal layer of endosperm is 
formed (figs. 23, 24). The large antipodal nucleus and those of 
the endosperm, like the original nucleus from which they sprung, 
are all large and contain very large central chromatin masses. 
This, in addition to their position at successive stages, points 
toward a common origin. The large antipodal nucleus continues 
to grow for some time and is at length very conspicuous. It is 
by far the largest in the ovule, and the very large deeply stain- 
ing chromatin mass may be seen even after the embryo has — 
reached a considerable size. | 
The nuclei in Potamogeton are all very peculiar, differing © 
from the ordinary type in having the chromatin mostly aggre- 
gated in a ball at the center of the cavity, instead of being dis- 
tributed on the linin network. Although several hundred slides 
were prepared, none contained spindles in the embryo-sac, and > 
consequently the sequence of the divisions had to be determined — 
by other means. The appearance of such a remarkable process 
of development, and one so different from those already 
described, although still strongly suggesting certain features of — 
the process found by recent investigators in related plants, 
made it very important that every step should be verified as far aS _ 
possible. For this reason the material was worked over several : 
