444 BOTANICAL GAZETTE [DECEMBER 
of the fusion of the polar nuclei (fig. 24). The endosperm nucleus 
which results from this fusion enlarges but does not divide, and 
soon degenerates along with the antipodals (fig. 25). 
The first division of the fertilized egg is transverse (fig. 16). 
The second division is in the micropylar cell and is also transverse, 
while the third (fig. 25) separates the.chalazal cell into two by a 
longitudinal wall. These two divisions frequently occur simul- 
taneously. Intermediate stages in the development of the embryo 
were not observed, but in the mature seed, which is of the usual 
orchidaceous type, it consists of a small, oval, undifferentiated 
mass of cells with no suspensor. 
Discussion 
Owing to the definite course of development in many of the 
animal eggs, the zoologists have been able to study some of the 
factors concerned. They have found in some cases that structures 
develop independently. In others some organs do not appear if 
certain parts are wanting, while in still other cases, as the lens of 
the amphibian eye (SPEMANN 14), structures, which at one time 
probably required the presence of another organ for their develop- 
ment, have during the course of evolution come to develop inde- 
pendently. 
The factors concerned in the development of plants have been 
studied much less than in the case of animals. This is perhaps 
due to the fact that most of the plants which show determinate 
development are inclosed, during their early stages, in the tissues 
of the parent. It is well known, however, that the form of a 
plant may be greatly affected by external conditions. A striking 
case is that of Stigeoclonium, in which, according to LIVINGSTON (8), 
the cells develop into a palmella stage or elongated filaments 
according to the osmotic strength of the nutrient solution. HARPER 
(6) in studying H ydrodictyon concluded that the shape of the net 
was due to the shape of the parent cell, while the axis of elongation of 
the individual cells was connected with the pressure exerted by 
neighboring cells upon each other. 
In Epipactis it is not evident why the nucleus of a megaspore 
should in some cases develop into the nuclei of a whole embryo sac, 
