448 BOTANICAL GAZETTE [DECEMBER 
produced in a variety of ways, but always fuse to form the endo- 
sperm nucleus, although this does not develop further. 
It would seem that even the final fate of the nuclei may depend 
largely on interacting conditions, for the synergids in those cases 
in which a sac is formed from four megaspores, as in the normal 
cases, are formed from the pair of nuclei arising from the transverse 
spindle. That the nuclei at this stage are equipotential is indicated 
by the occasional fertilization of one of the synergids (COULTER 
and CHAMBERLAIN 4). The structure of eggs, synergids, antipodals, 
etc., probably depends largely on the nature of the protoplasm 
of which they are constituted, and is of course widely different in 
different plants; but the part which any particular nucleus in 
Epipactis, and probably in other angiosperms, is to produce, as 
well as the general arrangement of the sac, apparently depends on 
the relation of the nucleus to other parts rather than upon any 
quality inherent in it. 
According to the above interpretation, the embryo sac in its 
early stages may be regarded as a system, all parts of which are 
equipotential, the fate of the different parts being connected with 
conditions external to them. The course of development in certain 
animal eggs is connected very largely with a stratification of the 
materials composing them, but in the early stages of many of these 
eggs a cell may develop into a whole embryo or some fraction of 
one, depending on whether or not it is separated from others. 
This dependence of the course of development of a cell on its rela- 
tion to conditions external to it, therefore, seems to be common to 
both plants and animals. 
The foregoing analysis, in so far as it goes, may be taken as 
indicating that the parts concerned act according to mechanical 
principles and do not need a vitalistic force to explain their behavior. 
This would seem to be true of any analysis which shows an orderly 
relation between an antecedent and consequent event, because for 
a thing to be mechanistic (this term being used in its widest sense) 
means simply that when the events are reduced to their simplest 
terms they take place in an orderly and predictable sequence. 
An analysis may bring to light new elemental laws of a different 
