1906] COOK—CUBAN NYMPHAEACEAE 381 
was the lower polar nucleus or one of the two endosperm nuclei resulting from 
the first division of the primary endosperm nucleus could not be determined. 
Whichever is the case, it never divides, and all the endosperm nuclei arise from 
the division of the other primary endosperm (or polar) nucleus. The endosperm 
is limited, there being usually no trace of cell formation. 
In discussing Zannichellia palustris he says: 
As in Naias, there is evident soon after fertilization a large nucleus just above 
the antipodal cells, which undergoes no division, but increases very much in size. 
This is more variable in size than in Naias; not infrequently it could not be 
detected in the later stages, and in several instances it looked as if it were under- 
going disintegration. 
Previous to this time, JoHNSON (12) had published his studies on 
Saururus cernuus, in which he described a division and behavior of 
the endosperm nucleus very similar to what I described for Nym- 
phaeaceae, and the formation of a nucellar tube sac intermediate 
between what I have described in this paper for NV ym phaea advena( ?) 
and Castalia ampla, JOHNSON (15) has since called attention to 
other genera of the Saururaceae (Anemiopsis and Houttuynia) which 
possess this character. 
In my previous discussion of this subject (7), I expressed the 
opinion that the physiological significance of this nucellar tube and 
nucleus presented a very interesting problem, which should be con- 
sidered in connection with the function of the antipodals. I called 
attention to the fact that in Ranunculaceae, Sparganium, and Vail- 
lantia the antipodals appeared to furnish nourishment for the embryo; 
that the peculiar haustorial development of the antipodals of Vail- 
lantia, the enlargement of the lower antipodal in Aster, the accumu- 
lation of endosperm in the antipodal region of Alyssum, and the 
large lower nucleus formed by a division of the endosperm nucleus 
(nucellar tube nucleus) in Sagittaria and the Nymphaeaceae showed 
a resemblance which I believed to indicate a similar physiological 
function. At about the same time Ikepa (11) published the results 
of his studies on the physiological functions of the antipodals, in 
Which he demonstrated by microchemical observations that the antip- 
odals of Liliaceae possessed very important physiological functions. 
JoHNson (14) also considers the antipodals of considerable 
Physiological importance in certain of the Piperaceae, in which he 
describes them as increasing in size and sometimes in number. It 
