334 
DUNCAN S. JOHNSON 
a whole and therefore is of no special significance in this case. It is 
surprising that in all of the Peperomias thus far studied there is no 
indication of the derivation of the bisporangiate stamen of this genus 
from the four-sporangiate stamen possessed by other genera of the 
Piperaceae, in common with the great bulk of other angiosperms. 
This loss of all traces of a four-sporangiate phase of the stamen, 
through which it seems evident that it must have passed in its phy- 
logeny, is an important fact. It will be well for us to have this fact 
in mind, when later we note the lack of any clear traces in the ontogeny 
of the phylogenetic origin of the peculiar type of embryo sac found in 
this genus. 
The nuclear phenomena occurring during the division of the 
sporogenous cells, and during the maturing of the spore-mother cell 
and its division into the spores, are essentially typical. The appear- 
ance of the chromatin net and nucleolus of the resting nuclei, formed 
during this development, remain practically constant up to the time of 
preparation for synapsis in the nucleus of the spore mother-cell. 
During the earlier divisions of the archesporial cells and their nuclei, 
divisions may not occur at the same time in the different cells 
(fig. 29). As the time for synapsis approaches, however, the nuclear 
changes go on at very nearly the same rate in all the mother cells 
of the same sporangium (fig. 32). The nuclear phases may differ 
slightly in rate of progress in the two sporangia of the same stamen. 
The chromatin net of the mother-cell, just before synapsis, is rather 
large-meshed in parts, and is roughened with many granules (fig. 36). 
The first evidences of the approach of synapsis are a loosening up of 
the net and an increase in size, apparently accompanied by a decrease 
in number, of the chromatin grains, until these become nearly as large 
and of very closely the same number as the chromosomes of the sporo- 
phyte (fig. 37). At this early phase of synapsis the chromatin frame- 
work is still distributed about the whole periphery of the nuclear cavity, 
most of it lying close to the wall. The nuclei of the stamen in the 
same spike, next older than the stamen whose nuclei show the arrange- 
ment just described, have each a tangled chromatin thread that is 
distinctly contracted away from the nuclear wall. In such a nucleus 
but few chromatin granules are visible. The completely contracted 
synaptic knot stage follows very soon after the stage just mentioned, 
and in it the nucleolus and all of the chromatin, except a few outlying 
granules are compacted into a rounded mass occupying less than a 
