Ig09] MCALLISTER—EMBRYO SAC OF SMILACINA 211 
His evidence lies in the fact of their being the product of the two reduc- 
tion divisions. Hence they must be “megaspore nuclei, to be recog- 
nized as such by their cytological history and structure.” The 16- 
nucleate embryo sacs of Peperomia pellucida are considered as result- 
ing from two divisions of each of these megaspore nuclei, and the 
8- and 16-nucleate embryo sacs, which JOHNSON reports as occurring 
in Peperomia hispidula, are regarded as formed by one division of 
each of the. four megaspore nuclei to form the 8-nucleate embryo 
sac, and by two divisions of each to form the ‘16-nucleate embryo 
sac. 
CouLTER is inclined to regard it as a fundamental law that the 
angiosperm embryo sac is formed from the mother cell by never 
more than five nuclear divisions, the reduction divisions and three 
divisions of a megaspore. He concludes that the large number of 
nuclei in the embryo sacs of Peperomia and Pandanus may originate 
by the participation of more than one spore in their organization. 
This implies, further, that any embryo sac formed from one spore 
which has sixteen or more nuclei can be considered as relatively 
primitive, since it would require more than five nuclear divisions to 
produce it from the mother cell. 
I do not see why the cases of the proliferation of antipodal nuclei 
should not be given more weight in the evidence. There is certainly 
an abundance of cases among the grasses and in the Ranunculaceae, 
as well as in other families, in which the innermost nuclei of the 
embryo sac are the product of more than five divisions previous to 
fertilization. It seems doubtful whether the embryo sac with more 
than eight nuclei can be explained on any such simple hypothesis, 
and it is to be remembered that, while there is apparently a physio- 
logical necessity back of the double division, there is nothing, ce 
STRASBURGER has noted, in the phylogeny of the angiosperms which 
would explain or give special significance to a fivefold division. 
Brown (1) reports that evanescent cell walls are formed separat- 
ing the first four nuclei of the embryo sac of Peperomia Sintensn 
and P. arijolia, in both of which the mother cell develops directly 
into the embryo sac. Though evidently convinced that in Peperemaes 
these first four nuclei are megaspores, BROWN seems to object to the 
adoption of this explanation for the lily type of embryo sac in general. 
