384 BOTANICAL GAZETTE [NOVEMBER 
From the facts in hand, the following statements seem to be 
justified: 
(1) Endosperm formation is not dependent upon the presence 
of a male nucleus. 
(2) Endosperm formation is not even dependent upon polar 
fusion. 
(3) Therefore, both of these fusions may be regarded as supple- 
mentary rather than determinative. 
(4) Endosperm formation does not even depend upon having 
been preceded by a reduction division. 
5) The fusions associated with endosperm formation do not 
represent a definite process, but are miscellaneous in number and 
order. é 
(6) The product of such fusions as do occur is merely an undif- 
ferentiated tissue, which practically continues the tissue of the 
gametophyte; that is, it is simply growth and not organization. 
Conclusions 
It seems evident that the egg has an-organization peculiar to 
itself. A male cell may fuse with any other cell in the sac, and the 
result is only endosperm; but occasionally such a fusion (as with 
a synergid or a polar) results in an embryo. This implies that, for 
- some reason, these ordinarily sterile cells have achieved the organl- 
zation of eggs. It is this possibility that makes them potential 
eggs; but in the ordinary embryo sac of angiosperms there is only 
one actual egg, which means only one cell capable of being fertilized 
in any real sense, and therefore capable of producing an embryo. 
Conditions in the embryo sac favor fusions of any free nuclel, 
in any number and of any origin. A male nucleus, perhaps, 1S 
more apt to enter into fusions than any other kind. 
A male nucleus entering into a fusion may or may not express 
itself as a carrier of hereditary characters. If it does express itself 
in this way, it is like injecting certain gamete tendencies into 4 
vegetative fusion; therefore, it is more probable that the male 
nucleus modifies somewhat the normal product than that the ant 
podal polar (a vegetative nucleus) modifies a normal product. In 
