1908] BROW N—PEPEROMIA 447 
mosomes. This is double the number found in the embryo sac, as 
will be described later. The chromosomes are small and short, and 
are therefore readily counted in a cross-section of a spindle at meta- 
phase. Fig. 2 shows such a section in the tapetum, while fig. 3 
represents a longitudinal view of a vegetative nucleus at a slightly 
older stage. Up to this time there certainly seems to have been 
no chromosome reduction, and nothing resembling megaspore forma- 
tion. : 
The single sporogenous cell (fig. 2) which is left after the cutting- 
off of the tapetum and which is to form the embryo sac is apparently a 
megaspore mother cell, as will be shown below. Its nucleus divides 
to two, four, eight, and finally in the mature sac to sixteen nuclei. 
The first division is heterotypic and takes place as follows. The 
resting nucleus (fig. 4) shows a meshwork of linin along which chro- 
matin granules are scattered. In the center of this meshwork is a 
large clear space containing a large nucleolus. After considerable 
growth the nucleus goes into synapsis. The meshwork contracts 
rapidly around, or to one side of, or even at some distance from, the 
nucleolus, into a mass in which very little detail can be made out. 
Fig. 5 represents an early stage of synapsis, while fig. 6 is probably 
older. No evidence of a fusion of spirems was seen either before 
or during the early stages of synapsis. 
At the end of synapsis the mass loosens up, and later appears in 
the form of a spirem, along which single granules are scattered at 
rather regular intervals (fig. 7). The spirem is apparently continuous 
and becomes loosely coiled, and the granules divide along the longi- 
tudinal axis of the spirem. A small portion of such a stage is shown 
in fig. 8. After this the spirem divides longitudinally and the two 
halves may diverge considerably in places (fig. 9), but later they come 
together again and all apparent traces of the division are lost. While 
this is taking place, the spirem is beginning to be arranged in loops 
(fig. 10), and is still apparently continuous, the loops being rounded 
at the ends. It does not seem possible that this appearance can have 
anything to do with the splitting just described. The looping be- 
comes more pronounced and the spirem segments transversely in 
such a way that.the loops give rise to chromosomes. There are eight 
of these, the haploid number, and they are apparently formed by the 
