368 Stephens . — The Embryo-Sac and Embryo of 
distinguishable, such a fusion of the second male nucleus with the primary 
endosperm nucleus is seen, 1 and there is no reason why it should not also 
take place in the Penaeaceae. 
If fertilization does not take place, the endosperm nucleus ordinarily 
disorganizes and disappears with the other contents of the sac. In one 
case, however, it was seen to divide before fertilization had occurred 
(Fig. 25). A similar formation of endosperm before fertilization (apparently 
the only clearly proven case among plants where fertilization normally 
occurs 2 ) has been reported as being occasionally seen in Ranunculus. 3 
Coulter suggests that the endosperm nucleus might here have been stimu- 
lated to divide through the influence exercised by the presence of the pollen- 
tube in the style being felt in the embryo-sac. The same explanation is 
possible in the case here figured, as pollen-tubes had already entered 
several other ovules in the same ovary. A like stimulus of the endosperm 
nucleus to division is seen in the Caprifig, 4 the exciting cause here being the 
presence of the egg of the Blastophagus wasp. Such cases of division of 
the primary endosperm nucleus before fertilization furnish a suggestive com- 
parison with the behaviour of the same nuclei in Welwitschia . 5 
Immediately after fertilization, the embryo-sac begins to enlarge rapidly 
at the expense of the surrounding nucellar tissue, until by the time the seed 
is ripe the nucellus has entirely disappeared. Even before fertilization the 
cells in the mid-line of the nucellus seem to part with some of their contents, 
losing in consequence their rounded outlines, and fitting together more 
compactly than the surrounding tissue (Fig. 24). If the ovule is not ferti- 
lized these axial cells simply collapse together under the pressure caused by 
fertilized ovules developing in the same loculus ; but if fertilization does 
take place, they rapidly disorganize (Fig. 26), and the embryo-sac, which is 
remarkably long and narrow, grows down into the cavity thus formed (cf. 
Figs. 27, 28 a , 30 <2, 35). The unfertilized cell-groups usually break down 
and disappear as soon as the embryo-sac begins to enlarge, but sometimes 
(very rarely) one or more may persist through the first few divisions of the 
embryo. As a rule, the endosperm nucleus begins to divide immediately 
after fertilization, or while fertilization is proceeding, and before the fertilized 
egg, 8-12 nuclei being present by the time the embryo is two-celled (cf. 
Fig. 27). (Fig. 29, however, if correctly interpreted, shows a case in which 
fusion of the primary endosperm nucleus is not yet complete, though fertiliza- 
tion has apparently taken place in the apical group and the other peripheral 
groups are already disorganizing.) I have not observed the stages of its first 
divisions, but later the endosperm nuclei have often been seen actively, and 
for the most part simultaneously, dividing (cf. Fig. 32). As might be 
expected from their multinuclear origin, they have a large number of chro- 
1 Modilewski, 1909, p. 23. 2 Coulter and Chamberlain, 1904, pp. 167, 170. 
3 Coulter, 1898, p. 83. 4 Du Sablon, 1908. 5 Pearson, 1909. 
