as A es‏ زر 
"eor. 
VAILLANTIA HISPIDA 18 
seems to be a barrier to any such migration, for the migrating mega- 
spore, so far as 1 have been able to determine, pushes its way to 
one side, as shown in Pl. 2, Fig. 1. It is reasonable to conclude 
that these capping cells are in a short time destroyed, for in later 
stages they are not usually seen. I have seen these cells once in 
one species of Galium, when the embryo-sac was mature. 
The nucleus of the migrating megaspore, after passing this 
barrier, comes to lie in the micropylar canal about midway its 
length, while the accompanying cytoplasm lies in a thin layer on 
the walls of the canal, and so forms a large vacuole at each 
end of the nucleus. The whole megaspore is, roughly speaking, 
hour-glass-shaped (Pl. 2, Fig. 1), and its nucleus lies at the con- 
striction. The enlarging megaspore encroaches laterally on the 
integument, which now for the first time is broken down. The 
endodermis gives way first, and is recognized a little later only by 
collapsed walls and the nuclei, which are the last elements of the 
cells to disintegrate. The persistence of nuclear substance under 
the action of digestive ferments suggests that the tissue which be- 
haves thus is undergoing digestion. This is the view taken by 
Mlle. Goldflus (A c.) in describing the behavior of the integu- 
mental tissue about the base of the embryo-sac in the Compositae 
the internal part of which is dissolved by “ferments secrétés prob- 
ablement par les cellules epithéliales.” 
At the constriction in the hour-glass-shaped megaspore the 
first division of the megaspore nucleus takes place, after which one 
of the daughter nuclei passes forward to the micropylar end of the 
cavity. When it reaches this point the second division takes 
place in both nuclei (Pl. 2, Fig. 2). There are now four. The 
egg apparatus arises in the usual fashion at the micropylar end 
(Pl. 2, Fig. 3, 4). In the antipodal region a more unusual state 
of affairs occurs. 
THE ANTIPODAL APPARATUS 
The division of the antipodal nuclei, now two in number, re- 
sults in four, one of which is, in the usual manner, concerned in 
the formation of the endosperm nucleus. This: passes forward to 
coalesce with the corresponding nucleus from the egg region. Of 
the three which remain, one enlarges considerably and migrates 
