214 _ BOTANICAL GAZETTE [SEPTEMBER 
after division of its nucleus gradually disintegrates; the lower cell 
also gives rise to two nuclei, one of which, the lower, becomes 
a mere clump of chromatin, while the other divides twice and 
the four resulting nuclei arrange themselves as in Clintonia and 
the above named Onagraceae. Though WENT does not discuss the 
theoretical value of the first four nuclei derived from the mother 
cell, it seems clear they represent megaspores, of which that next 
the innermost is the fertile one. Thus two megaspores appear 
in the upper cell, and two in the lower cell which becomes the 
embryo sac. The development of one of the middle megaspores, 
although uncommon, is not unknown. It has been seen in Acacia 
and Eriobotrya (GUIGNARD 1881, 1882), Trapella (OLIVER 1888), 
some of the Araliaceae (DucAMP 1902), and Asclepias (FRYE 
1902); to this list, which is taken from CouLTER and CHAMBER- 
LAIN’S Angiosperms, may be added Vaillantia and Collipeltis 
(LLoyD 1902). 
In the examples thus far reviewed, the four functional nuclei of 
the embryo sac are the direct derivatives of one of four megaspores. 
Some cases of a different nature remain for consideration. In 
Limnocharis (HALL 1902) tetrads are not formed; the first two 
nuclei of the mother cell place themselves at opposite poles, and 
while the upper gives rise to the egg apparatus and a polar nucleus, 
the lower remains undivided. Nearly similar is Helosis, but in 
this case the primary antipodal nucleus soon degenerates (CHODAT 
and BERNARD 1900; I have not been able to consult the original 
paper). Cypripedium (Pace 11) furnishes another example of a 
tetranucleate embryo sac. In this plant the mother cell divides 
once and the lower of the two daughter cells becomes the embryo 
sac, its nucleus undergoing two divisions. Miss Pace interprets 
the first two nuclei of the embryo sac as megaspore nuclei. Thus 
according’to this view the embryo sac of Cypripedium is compound, 
being the product of two megaspore nuclei. 
Miss PAce extended the conception of a compound embryo 
sac to Lilium, in which four megaspores are thought to function, 
and more recently CouLTER (4) has extended it to all those cases 
in which tetrad formation is apparently suppressed, and espe- 
cially to the 16-nucleate embryo sac of Peperomia and the like. 
