1909] BRIEFER ARTICLES 57 
destroyed in the preparation of the material. The writer has observed it 
repeatedly in Castalia odorata and in Nymphaea advena, where it can be 
seen without great difficulty; also in Cabomba piauhiensis and in Brasenia 
purpurea, where it is very difficult to demonstrate. York noted it in Ne- 
lumbo lutea, but according to his observations one of these two cells again 
divides, thus forming three superposed cells, each of which is active in 
the formation of the endosperm. However, he did not observe the formation 
of the large vesicular cell, which is so characteristic of the other genera of 
this family. My recent studies, which will be referred to later in this paper, 
do not fully agree with those of Yorx in regard to the vesicular cell. 
The daughter nucleus in the micropylar end of the sac always divides to 
form the endosperm. The formation of the nucellar tube, into which the 
other nucleus usually and possibly always passes, is subject to considerable 
variation in the different species. In the genus Nymphaea it is a cylindrical 
tube, which extends almost the entire length of the ovule (fig. 1). In Cas- 
talia it may be in the form of a tube which is practically the same as in Nym- 
phaea, but it is usually somewhat shorter and narrower. In some of my 
recent studies I have found individuals in which there were well-formed 
tubes extending only about one-half the length of the ovule (jig. 2), while 
in other individuals it extended not more than one-third the length of the 
ovule and tapered to a point (fig. 3). In Castalia ampla, as previously 
reported by me, it develops into a short, sac-like structure (fig. 4), which 
is eventually overgrown by the micropylar part of the sac (jig. 5). In 
Brasenia purpurea and Cabomba piauhiensis, the tube is very long, extend- 
ing almost the entire length of the ovule, and very small (jig. 6). In Ne- 
lumbo lutea, according to these studies, the tube is very irregular‘*in form 
and extends about two-thirds the length of the ovule (jig. 7). 
In these recent studies one individual of Nelumbo lutea was observed 
in which two sacs were formed one below the other in the main axis of the 
ovule (fig. 8). The antipodal ends were in contact, and the sac next to 
the micropyle was about four times as large as the other. It showed a 
fully developed egg-apparatus, and the two polar nuclei were in contact, 
but the antipodals had disintegrated. The small sac which was farthest 
from the micropyle showed the egg, one synergid, the three antipodals, 
and was filled with the characteristic endosperm. In the antipodal end of 
the sac was a small, deeply colored mass, which was possibly the remains of 
the disintegrating tube nucleus (fig. 8#). This extra sac was evidently 
derived from one of the sister megaspores. York observed many cases in 
__ Which extra, imperfect sacs were produced, and expressed the opinion that 
__ they were from sister megaspores rather than from independent mega- 
