( 384 ) 
lose sight of the fact that such conditions exist in many of 
the lower angiosperms. 
In the division of the primary sporogenous cell four mega- 
spores are usually formed, and the condition in this family is 
therefore in harmony with the almost universal habit in the 
Sympetalae. In all the cases where it was observed the 
tetrad was a straight row of cells of which it was always the 
one at the chalazal end of the series which became functional, 
the only apparent reason for the selection of this cell being 
its more advantageous situation with reference to the source 
of nutritive materials which come through the chalaza, an 
idea which Koernicke has also expressed.” In all cases 
the spore enlarges considerably and then proceeds to the 
regular series of divisions, which usually gives rise to but 
eight nuclei from which the egg-apparatus and other struc- 
tures are differentiated. Very few mitoses were observed in 
all the preparations examined and no generalizations can 
be made concerning conditions upon which such facts are 
supposed to throw light. Where the divisions were observed 
they were simultaneous within the same sac, and there was 
no evidence of the most evanescent kind of a cell-plate. In 
the formation of the egg-apparatus and the development of 
its particular features we recognize those characters so com- 
mon among the dicotyledons and the Sympetalae in particu- 
lar. In fact in all essential features the embryo-sac is similar 
to many others, from various sources and cycles of affinity. 
The polar nuclei in many cases have been observed to unite 
before any evidence of fertilization was apparent. While 
neither of the polar nuclei was observed to lie in contact with 
the egg-cell their position was uniformly close to it. The 
lower of the polar nuclei always migrates to the vicinity of 
the egg-apparatus, where it unites with the upper one; 
but no migration of the fusion nucleus occurs as observed by 
Ikeda* in Zrccyrtes hirta. 
Among the Sympetalae there is a tendency toward the re- 
tention of the antipodals, and even in some cases toward the 
formation of tissue by their proliferation. But such is not 
