1912] STEVENS—HETEROSTYLOUS PLANTS 291 
easier by a change which takes place in the cell at this time, and 
by which postsynaptic stages are clearly distinguished from pre- 
synaptic stages. During the contracted condition of the chromatic 
mass the pollen mother cell becomes rounded and takes on a more 
spherical form. In figs. 3, 4, and 5, the nuclei are apparently in 
much the same stage, but the cells show a progressive ‘rounding 
off.” Such a change in the shape of the cell during synapsis has 
been observed by STRASBURGER, ALLEN (1, fig. 19), Davis (12, 
p. 634), and others, and appears in the figures of many workers 
who make no particular mention of it. 
The synaptic contraction seems to take place by a drawing 
together, at one side of the nucleus, of the whole nuclear reticulum, 
usually but not always including the nucleolus (figs. 3 and 5). On 
careful examination this mass appears to consist, at least in part, 
of delicate threads; but no evidence of any pairing of these threads, 
such as has been described by some investigators, could be obtained. 
The threads of the nuclear. reticulum in the buckwheat are so 
delicate, however, that it would be extremely difficult to demon- 
strate any such condition even if it occurred; and the behavior 
of the chromatin at later stages makes it seem probable that a pair- 
ing has actually taken place at this stage. 
The amount of chromatin staining material is very markedly 
increased during synapsis. Before the contraction, the greater 
part of the reticulum does not take the chromatin stain; but the 
nuclear mass comes out of the contracted condition as a series of 
rather thin loops (fig. 7), each of which consists, apparently, of a 
single thread which takes the chromatin stain uniformly throughout 
- itslength. This thread is granular in appearance and varies some- 
what in thickness; but no alternation of chromatic and achromatic 
material, such as has been described by some writers at this stage, 
could be made out. The number of loops is rather inconstant, 
but is generally greater than the number of gemini. 
These loops gradually shorten and thicken (fig. 8) and become 
more dense and uniform in appearance. At the time of greatest 
thickness they undergo a longitudinal split (fig. 9), thus giving rise 
to a series of paired chromatic threads from which the gemini 
are apparently formed by the continued gathering together of the 
