EMBRYO-SAC OF GYMNADENIA CONOPSEA. 
O 
more fluid material for some time these masses are ill 
defined, and they never form a distinct cell-wall, but, finally, 
from the presence of a nucleus and well-marked contour, w^e 
recognise in them naked cells in a protoplasmic matrix 
Each of these soon divides into four smaller but otherwise 
similar nucleated masses, by two divisions in planes crossing 
at right angles. In some cases, at any rate, this occurs 
simultaneously in both, but in others one or both divide first 
into two, and then again each daughter mass into two others 
(cf. figs. 8—12). 
That this is the final result there can be no doubt, but 
whether one is to suspect variations in the divisions as due 
to manipulation, conditions of vigour, or peculiarities of the 
species, must remain at present undecided. The important 
fact is that eight nucleated masses of protoplasm result, 
more or less isolated and complete, in groups of four at each 
end of the sac (figs. 12, 13) ; this I consider demonstrated 
by the specimen figured at 12, though the division of the 
lower mass does not always appear to be completed.^ 
The relations of the products of division in the fore part 
of the sac are, however, remarkably constant and surprising. 
Two of them become elongated and packed close into 
the top of the embryo-sac, as the Gehiilfinnen or “ Sy- 
nergidae ” of Strasburger ; while one enlarges and rounds 
off* as the egg-cell or oosphere,^^ and becomes suspended 
laterally at the base of the “ Synergidae ” in the cavity of 
the sac. The fourth mass also rounds off, and falls freely 
into the lining protoplasm of the sac. All these masses have 
acquired nuclei, that of the egg-cell being especially large 
and bright, but no trace of a cell-wall appears around any of 
them. 
As to the fate of the products of division in the posterior 
end, I can say very little positively. That four masses com- 
mence to form appears certain, but only in a few cases have 
they become completed ; as a rule, I find a mass of protoplasm 
in this place with a variable number of nuclei in it, but in 
some cases (fig. 13) four masses occur. 
It is to be regretted that this point has not been more 
successfully dealt with, and also that the exact origin of the 
* Many facts suggest that this division is of the same order as the two 
preceding divisions of the “en»bryo-sac motlier-cell.” If so, we must con- 
sider that cell as suffering division into four, the last division wall being a 
very weak and diffluent one. 
’ The lower group may apparently suffer less complete division in other 
plants also. 
