and Affinities of Trapella . 89 
At the time when the two proximal cap-cells (c 1 and c 2 ) begin 
' to dwindle, the embryo-sac goes through the stages which 
precede fertilization. The single nucleus of the embryo-sac 
divides into two (Fig. 27 ), and each of these gives rise in the 
normal manner to the egg-apparatus, antipodals, &c. In Fig. 28 
the embryo-sac is ready for fertilization. Its upper (micro- 
pylar) end has become much widened and its apex pointed. 
A line (in) representing the micropyle is seen in Fig. 27 , and 
in Fig. 28 this is occupied by the pollen-tube (p. t.) The 
nuclei constituting the group of antipodals (ant.) are of small 
account and shortly dwindle. 
It must be observed that in this stage the tip of the apical 
cap-cell (c 3 ) has become pointed, and that it gradually elon- 
gates downwards. A later stage is given in Fig. 29 : this 
however is taken from the lower ovule. The apical cap-cell 
here almost equals the embryo-sac in length. In the embryo- 
sac itself I have only drawn one nucleus, as the others were 
not distinguishable. The stage is given to help to fill a 
gap in the history of the upper ovule. It seems, however, 
certain that this cap-cell steadily elongates. 
From this point the lower ovule gradually retrogrades, and 
hangs, in older stages, as a shrivelled remnant, which may be 
found with care even in ripe fruits, as already noted. 
The upper ovule now begins to elongate very rapidly so as 
to extend right down into the lower region of the loculus. 
That active division of the cells of the nucellus and integument 
is taking place may be seen from the arrangement in rows of 
the flat cells on either side of the embryo-sac in Figures 26, 
27 . The stage next following, in which the egg-cell undergoes 
its first segmentation and endosperm begins to form, I have 
not been fortunate enough to secure. My youngest fruit, 
though younger than that given in Fig. 12, where the spines 
are just showing prominently, has a much elongated ovule, in 
which profound changes have already occurred. The changes 
must be very rapidly passed through at this point, since in the 
material at hand I have been able to obtain numerous 
preparations of the stages both before and after the hiatus 
