210 BOTANICAL GAZETTE [SEPTEMBER 
in Oenone and Mourera a four-celled embryo sac is formed, similar 
to that in Helosis. The mother cell following synapsis divides to 
form two daughter cells, the upper degenerating. The lower daughter 
cell (which he calls the “megaspore’’) divides, and the innermost 
nucleus shrinks immediately to a shapeless mass of chromatin that 
is visible for a long time. This inner nucleus, though one of the 
first four nuclei formed by the division of the embryo sac mother 
cell, is interpreted by WENT as a nucleus of the embryo sac and not 
a spore. The nucleus remaining from the second division divides 
twice to form the four-celled embryo sac. The lower nucleus of this 
embryo sac degenerates, leaving only the three nuclei which form the 
egg apparatus. WENT notes that his results support the theory of 
Porscu, except that the ventral canal cell is on the wrong side of the 
cog. 
ERNST (10) has expressed the opinion that the 16-nucleate 
embryo sac of Gunnera, as described by him, consists of two of the 
archegonia of PorscH in the chalazal end of the embryo sac and one 
in the micropylar end. Four nuclei in the central region of the 
embryo sac fail to form an archegonium, and unite with a polar 
nucleus from each archegonium to form the endosperm nucleus. 
STEPHENS (23), in a preliminary note on certain Penaeactaé, 
reports that the sixteen nuclei of the embryo sac become divided into 
four groups, which lie at some distance from one another against the 
wall of the embryo sac. Three nuclei out of each group of four 
organize what appears to be an egg apparatus, while one nucleus 
from each group acts as a polar nucleus, and the four unite to form 
the endosperm nucleus. On Porscu’s theory these four groups 
would represent four archegonia. 
These examples seem to indicate a tendency of the nuclei of the 
angiosperm embryo sac to form groups of four. The evidence in 
favor of the archegoniate character of these groups, however, seceia 
to me to be insufficient as yet. The relationship between the angio” 
sperms and the gymnosperms is so remote that a comparison of the 
embryo sacs of the two groups with a view to homologizing sts 
groups of nuclei associated with the egg is very difficult. ae 
CouLTer (8) further defends the idea that the first four nucle 0" 
the embryo sac of the lily are megaspores, or “at least their nuclel- 
