206 BOTANICAL GAZETTE: [SEPTEMBER 
first and second divisions of such a megaspore to form a four-celled 
embryo sac. We must conclude, therefore, that the walls between the 
four megaspores break down, and their nuclei become the first four 
nuclei of the embryo sac. 
In the succeeding stages there is a continuous growth of the nuclei 
resulting from the reduction divisions and of the embryo sac formed 
from these nuclei. Very soon after the cell walls have disappeared 
from between the reduction nuclei, vacuoles begin to appear in the 
young embryo sac. Figs. 15, 16, 17 illustrate the gradual enlarge- 
ment of the cell containing the four nuclei, and the appearance of 
vacuoles in the cytoplasm. The vacuolization more than keeps 
pace with the enlargement of the embryo sac, and finally the separate 
vacuoles unite to form one large centrally located vacuole. It is at 
about this stage of growth that the third division takes place, forming 
the eight nuclei of the complete embryo sac (fig. 17). The embryo 
sac still continues to enlarge. The nuclei remain unchanged for a 
relatively long period. The polar nuclei come together and lie in 
contact during this period of quiescence, but are seen to be fused 
before the cells in the micropylar region show any signs of differentia- 
tion to form the egg apparatus. Not until nearly time for pollination 
is there any rearrangement of these cells, which are to form the egg 
apparatus. 
When finally differentiated ‘the synergids are pear-shaped and 
faintly striated. The egg is somewhat larger than the synergids, and 
usually has a large vacuole in the basal region. It stains less heavily 
than the synergids. I shall not here discuss the formation of the 
cell boundaries of the synergids and the egg. The antipodal ney 
are smaller than the others, and usually occupy a constricted regio” 
at the lower end of the embryo sac. These antipodal nuclei at 
stage often number more than three, and in such cases are usually 
separated by division walls. : 
The most natural interpretation of the phenomena just described 
is that the first four cells formed by the division of the embry° - 
mother cell are megaspores, and that these four spores jointly com 
bine to form one embryo sac. It would seem that this assump? 
is the only one possible, for before the division membranes disap 2: 
the four cells conform to all the criteria for spores in other similar 
