STUDIES IN THE DEVELOPMENT OF THE PIPERACEAE 737 
division of the microspore nucleus. The ripe microspore is uni- 
cellular and binucleate. The primary archesporial cell of the 
ovule cuts off a single parietal or tapetal cell above. Then the 
lower half, which is perhaps to be considered a megaspore mother- 
cell, gives rise immediately to an eight-nucleate embryo-sac of 
the usual angiospermous type. 
Fertilization and endosperm-formation take place in the normal 
way. About 100 free, peripheral nuclei are formed before cell 
walls appear in the endosperm. The antipodals multiply to 
100 or more. In the ripe seed the embryo consists of a globular 
undifferentiated mass of about 500 cells. The endosperm is 
of 150 cells in median longitudinal section. Its cells contain 
little stored carbohydrate. The numerous antipodals persist 
in the seed, but seem to have little stored food-material in them. 
The abundant and starchy perisperm is the chief storage tissue. 
The fruit is immersed in the axis, but aside from the differences 
in structure connected with this fact, it resembles the fruit of 
Piper medium. 
The most striking peculiarity of this species is the extreme 
variability in the development of the microsporangia and mega- 
sporangia on different spikes or in different flowers of the same 
spike. The number of microsporangia in a stamen e.g., may vary 
from none to four, and the extent of a single sporangium may be 
such as to fill one-quarter of the anther, or, in others, as much 
as three-quarters. The number and relative extent of the sporo- 
genous masses in the stamen is constant from the time of their 
initiation. There is no breaking down later of sterile septa to 
throw two sporogenous masses into one. There is no secondary 
fusion or confluence of sporangia, nor is there any evidence of 
the abortion or suppression of sporogenous tissues once initi- 
ated, unless it be found in the possible infertility of the well- 
formed pollen of some of the ripe anthers. 
The cause of the differences in development of the sporogenous 
tissue seems not to be connected in any way with the space rela- 
tions of the flowers on the spike, since any type may occur at 
any point on the spike, i.e., at base, middle or tip. The real cause 
is probably to be sought in those factors, internal or external, 
THE JOURNAL OF EXPERIMENTAl ZOOLOGY, VOI . 9, NO. 4. 
