380 BOTANICAL GAZETTE [NOVEMBER 
(fig. 9a). The tube is flat (jig. 9b) instead of cylindrical as in the case 
of the other species. This tube is usually completed just before 
the appearance of the cotyledonary ridge of the embryo. The nucleus 
was observed in the antipodal end of this tube but apparently dis- 
integrated, and did not pass to the opposite end of the tube as in the 
case of the other species. 
In Brasenia purpurea and Cabomba piauhiensis, the nucellar tube 
is long but very small, There is a rapid disintegration of the nucellar 
tissue to form the tube before the tube nucleus begins its passage. 
The tube nucleus gradually disintegrates, and has entirely disappeared 
before it reaches the chalazal end of the tube (jigs. 7, 10, 11); the 
chalazal end of the tube is considerably enlarged, is not so definite 
in outline as in Nymphaea advena, and contains fragments of the 
cell walls (fig. 17). 
In my previous discussion (7) I called attention to the fact that 
the behavior of the endosperm nucleus is similar to that of Sagittaria 
variabilis as described by SCHAFFNER (25), except that in S . variabilis 
there is no nucellar tube elongation; also, that judging from the 
studies of WrEGAND (27) and HoLrerty (10) on Potamogeton it is 
possible that a similar condition may be found in that genus. 
Since that time, Hai (9), in his studies on Limnocharis emarginata, 
in which he finds a single polar nucleus, says: 
The upper polar nucleus, when it has approached the antipodal end of the 
sac, divides transversely. The lower daughter nucleus remains in the position 
of its formation, being cut off by a wall across the sac and forming a large cell 
which does not divide further, but finally disappears through the encroachment 
of the endosperm... . . The upper daughter nucleus travels back towards 
the egg apparatus, and by its further division forms the endosperm. 
STRASBURGER (23) describes a similar action of the endosperm 
nucleus of Ceratophyllum submersum, in which, after the first division, 
the nucleus in the antipodal end of the sac does not divide, while the 
nucleus of the micropylar end forms the endosperm. : 
CAMPBELL (1) in his discussion of Naias flexilis and Zannichellia 
palustris describes a condition which further study may prove to 
be similar to the Nymphaeaceae. In discussing Naias he says: 
A peculiarity noted, which was also observed in Zannichellia, was the presence 
of a single large nucleus close to the antipodals, which was conspicuous at 2n 
early period and behaved much like the nucleus of the suspensor. Whether this 
