386 BOTANICAL GAZETTE [NOVEMBER 
structing such an embryo from serial sections, one might readily take it for a 
dicotyl. It will be evident, however, from a comparison of the figures that the 
Castalia embryo represents only the extreme of the lobing shown in Nelumbo 
and Nymphaea. 
The formation of the suspensor is entirely different from what 
I observed in C. odorata, but corresponds with Conarn’s (5, 6) 
observations. The suspensor disappears soon after the formation 
of the cotyledonary ridge. 
The development of the embryos of Brasenia purpurea and 
Cabomba piauhiensis is practically the same, but since C. piauhiensis 
is much more easily sectioned and furnished much better prepara- 
tions, most of the drawings were made from it, The fertilized egg 
divides by the formation of cross walls and produces a proembryo 
of three or four cells in linear arrangement (figs. 32, 33a, 34, 35). The 
terminal cell then divides, forming a quadrant (jig. 33a), then it 
forms the octant, and then a large spherical embryo supported by a 
short suspensor of two or three cells which usually divide longitudi- 
nally (figs. 34, 35, 36). As the embryo increases in size, it becomes 
more or less flattened against the walls of the sac and develops the 
cotyledonary ridge (jig. 36, x) similar to the two genera just described, 
except that this ridge extends almost entirely around the plumule, 
thus forming a pit with the plumule in the center. The dicotyledonous 
character produced by the development of the cotyledonary lobes 
appears very early. Only by the most careful examination of the 
intermediate stages at the time of the first appearance of the cotyle- 
dons, and by the most careful cross sections was it possible to demon- 
strate the common origin of these two cotyledonary lobes. They 
develop very early and the edges and tips come together, thus enclos- 
ing the plumule in a short, hollow cone (fig. 37). After this the 
development of the embryo is a mere increase in size (figs. 38, 39), 
accompanied by the modification of the endosperm previously referred 
to. The suspensor persists until the embryo is almost mature and 
then disintegrates. 
It will be noted that the young embryo of Nymphaea advena (?) 
is similar to the embryos of Lysichiton kamtschatense as described by 
CAMPBELL (2); and of Ceratophyllum submersum as described by 
STRASBURGER (23). The young embryos of Castalia ampla, C. 
