1903] EMBRYO SAC OF CASUARINA | 107 
of the protoplasm of the sac, strongly suggesting the separation 
of these cells by a wall. Comparing now fg. rg, C, with fig. 15, 
a slightly older sac, the antipodals in the latter certainly cannot 
be questioned. The whole sac is typical enough of angiosperms 
to serve as a conventional prefertilization sac. However, as has 
been remarked, fig. 76 is more typical of the sac of C. stricta. In 
the older sacs (figs. 16, 77, 20) these antipodals resemble the 
unabsorbed megaspores very much, and could easily be mistaken 
forthem. This might account for their reported absence, unless 
other species differ in this respect from C. stricta. 
Sometimes, however, no antipodals were found (fig. 27), and 
the reason for this may be considered. Some of the sacs develop 
long antipodal prolongations (figs. r4, 77, 20), and these do not 
always appear at the same time. ig. 77, a shows one already 
under way when the sac is in the 2-celled stage, and 4-celled 
stages were also found showing the beginning of antipodal pro- 
longations. ig. 13 may be developing one at a, although that 
cannot be said with certainty. Ina case like fig. rz, when the 
sac reaches the 8-celled stage and the antipodals settle at some 
point, where will they likely be found? If they seek the antip- 
odal end, they will likely slip into the antipodal prolongation 
and perhaps be found somewhere within that. At least such a 
thing seems quite possible. The antipodal prolongations, how- 
ever, are long tortuous tubes, very much entangled with their 
fellows. It is almost impossible to trace a definite one to its end 
with certainty. Much more is it difficult to determine the num- 
ber of nuclei within, when above and below it are scattered 
unresorbed undeveloped megaspores. There are nuclei within 
these prolongations in some cases in which antipodals are want- 
ing, but I was not certain of the number in any case. However, 
in one case there were three cells (fig. 14, D, h, 2, j) at various 
distances down the tube. If in this sac the polars had already 
united, all the nuclei are accounted for. Sometimes, however, 
the antipodal prolongations are not formed until late in the his- 
tory of the sac (figs. zg C, 75). If it occurs after the antipodals 
have become inclosed by walls and perhaps adhere to the walls 
of the sac (fig. 16, a), their presence in the body of the sac, 
