92 
as described by the author (Pam 1915) for Zanacetum. These 
demand special investigation since they form connecting links 
between the so called normal and the sixteen nucleate embryosac 
of Compositae. 
The embryosac seems to remain for a considerable period in 
the two-nucleate stage (fig. 4). During this time a active elon- 
gation of the sac has set in, the upper vacuole pressing the 
two embryosac nuclei apart. 
There is still a pronounced difference in size between these — 
two nuclei, in spite of both having increased considerably in 
\ volume during this period of 
growth. The degenerating basal - 
megaspore are separated from 
the two embryosac nuclei by a 
vacuole. 
The divisions following still 
maintain the difference between 
the micropylar and chalazal 
groups of nuclei in the embryo 
sac. The micropylar pair of nuclei 
imbedded in densely stained 
plasma is composed of distinctly 
larger nuclei than the chalazal 
one (fig. 5). Both pairs are sepa- 
rated by a vacuole that now 
occupies a central position in the 
embryosac. 
The next step in the develop- 
ment results in an eight nucleate 
rey embryosac of the type generally 
bryome with four vie’ eight nuclei; the f0UNd in the Composita. In fig. 6 
— ms —_ still visible. the partition walls between -the 
oP i constituting elements of the fe- 
male gametophyte are just being laid down, quite in the usual 
way. Thus in the dense micropylar plasma the nuclei of the 
synergids — these always characteristically contain 2 or 3 
nucleoli — of the eggcell, and of the upper polar are found. 
- . oad 
De a 
SAS 
“2, 
coat feet 
eee 
uy 
oF 
i 
we 
Ls 
; 
oe 
