556 Davis . — Cytological Studies on Oenothera . /. 
clearly all of its parts. Sometimes, however, an occasional small nucleolus 
will be left free in the nuclear cavity (Fig. 13). There are usually at the 
edge of the synaptic knot a number of threads in the form of loops that 
extend often to the periphery of the nucleus (Fig. 14), but more frequently 
lie at some distance from the nuclear membrane (Fig. 15). It is clear that 
the process of contraction is a gradual one, and that some of the threads are 
drawn in more slowly than others and consequently remain free from 
the bulk of the contracted structure. Even in advanced stages of synapsis 
there are frequently present some extremely delicate threads, which later 
disappear, lying close to the nuclear membrane (Figs. 13, 15, and 19). 
Differential staining of the chromatic threads, especially of such free 
loops as are favourable for study, shows them at certain stages to be 
composed of denser, deeply staining regions arranged like a string of beads 
(Fig. 14). These are apparently the oft-described chromomeres comparable 
to similar structures on the early spirem of the vegetative nucleus (Fig. 4). 
The total length of the thread system in the early stages of synapsis is 
certainly very much greater than that of the early spirem of the vegetative 
nucleus, and apparently even greater than might be indicated by the propor- 
tionate increase in the size of the nucleus (compare Figs. 3 and 4 with Figs. 
13-15). It is impossible to say whether the coiled and twisted elements of 
the synaptic knot are organically parts of a continuous spirem. They are 
certainly united to one another at many points, but this union may be 
merely an adherence which does not signify the branching of a true chromatic 
thread. 
As would be expected of such complex structures as the various stages 
of the contracted synaptic knot it is not difficult to find threads that run 
closely parallel with one another for considerable distances. Furthermore, 
as stated above, the threads are certainly united at points (Figs. 14, 15, 16, 
and 17), but they form such various angles and lie in such complicated coils 
that a clear understanding of their intimate relations to one another 
was impossible with the technique employed. Further study employing 
methods of fixation suggested by Overton (’ 09 ) may give clearer information 
on the conditions. The writer has, however, so far seen no evidence that these 
parallel threads ever fuse for considerable lengths, or that there is present in 
the nucleus two independent spirems (maternal and paternal) which might 
be assumed to unite with one another side by side. The further history 
of the threads in relation to the bivalent chromosomes also gives no 
support to such an interpretation. Neither has the writer observed any 
stages that would establish the splitting of a spirem, such as might indicate 
a premature fission related to the lengthwise division of the chromosomes in 
the homotypic mitosis. 
The contraction of the chromatic material during synapsis is apparently 
caused by the gradual shortening and thickening of the threads which later 
