1911] BROWN & SHARP—EPIPACTIS 449 
kind from any that we know at present, but in so far as they 
are laws of an orderly sequence, they will be as good a mechanical 
explanation as any other law, for a law can only state the sequence, 
and it is outside the realm of science to explain why one event fol- 
lows another. Any vitalistic explanation must therefore be either 
outside ‘and supplementary to science or contrary to the funda- 
mental postulate of all science, namely, that the same antecedent 
conditions are always followed by the same consequent ones. 
If we compare the development of the angiosperm embryo sac 
with that of the gymnosperms, we find in the early stages a strik- 
ing similarity between those of the gymnosperms and the sixteen- 
nucleate sacs of the angiosperms. In both cases the nuclei are 
fairly numerous, evenly distributed in the cytoplasm, and do not 
Show a polar arrangement. This similarity, however, is probably 
derived and not primitive in the case of the sixteen-nucleate sacs, 
for some of these, at least, are derived from four megaspores. There 
would appear to be in the gymnosperm embryo sac nothing similar 
to the striking polarity shown by those of most angiosperms, but 
that the same factors are at work is perhaps indicated by the 
elongated shape of the embryo sacs of many of the gymnosperms, 
as well as the tendency toward a reduction in the number of nuclei, 
and the presence of a large central vacuole. Likewise the presence 
in the early stages of the gymnosperm embryo sac of free nuclei 
Surrounded by a cellular region may foreshadow the free polar 
nuclei of the angiosperms. Porsc# (12), in an excellent discussion 
of the phylogeny of the angiosperm embryo sac, has attempted 
to point out a similarity between the archegonia of the gymno- 
sperms and the two polar groups in the angiosperms. When 
we remember, however, that in those gymnosperms which have 
archegonia they are initiated in a cellular phase and the polar 
groups of the angiosperms in a non-cellular one, it would seem that 
any similarity between the development, final structure, or factors 
concerned must be rather superficial. It would probably be better 
to regard the structure of the angiosperm embryo sac as the result 
_ot new physiological conditions which have arisen in connection 
with the reduction of its size and the number of its nuclei. 
