7 B4 Campbell .— 77 ^ Embryo-sac of Pandanus. 
tically impossible to determine their number. The two nuclei resulting 
from this first division were not seen, and the next stage met with showed 
four large endosperm nuclei, one of which is shown in section in Fig. 29, a , 
PL LX. These nuclei were of about the same size as the primary endosperm 
nucleus and showed much the same structure, containing several large 
nucleoli and a conspicuous reticulum. 
With the successive divisions of the endosperm nuclei there is a 
marked reduction in their size, and this reduction in size is accompanied 
by a reduced number of chromosomes, as may be seen by comparing 
Fig. 28, b, with Fig. 38, a. The former nuclear spindle was from a nucleus 
of about the same size as that shown in Fig. 29. The free endosperm 
nuclei are evenly distributed through the cytoplasm lining the embryo-sac, 
and this cytoplasmic layer increases a good deal in thickness before the 
formation of cell-walls in the endosperm begins. At the time the first 
walls are formed in the endosperm, the nuclei are still further diminished 
in size (Fig. 29, a), and presumably the number of chromosomes is corre¬ 
spondingly less than in the earlier and larger free endosperm nuclei. 
The Embryo. 
The embryo in Pandanus remains very small even in the ripe seed, 
and in the largest ovules that were sectioned, which had attained a length 
of more than a centimetre, the embryo was still quite undifferentiated. 
The early divisions take place slowly and the embryo remains unicellular 
for a long time after fertilization (Fig. 33). The first division, in some 
cases at least, is transverse and separates a short suspensor cell from a 
larger terminal cell (Fig. 34), but the number of young embryos examined 
was too small to make it certain whether the first divisions are always 
the same. Text-fig. 2, A, B, shows a four-celled embryo with probably a 
one-celled suspensor (j), although it is possible that this basal cell may 
belong to the nucellus. 
The oldest embryos seen are shown in Text-fig. 2, C, D. These were 
nearly oval bodies without any definite suspensor. They showed no signs 
of the permanent organs of the young plant, and it is therefore impossible 
to say what is the method of origin of the stem apex and the other organs 
of the seedling. In order to follow out the further history of the embryo 
it would probably be necessary-to germinate the seeds, as even in the ripe 
seed the embryo is quite rudimentary. 
In the development of the embryo Pandanus differs very much from 
Sparganium , where the embryo in the ripe seed is very large and all the 
organs perfectly developed. The earlier stages in Pandanus are not very 
unlike the corresponding ones in Sparganium (compare for example Fig. 37 
of my paper on Sparganium with Text-fig. 2, c). 
