945 
Oenothera Lamarckiana and O. gigas. 
extending over a considerable part of the long period leading up to synapsis. 
It involves the gradual thickening of what are at first relatively few very 
delicate threads (Fig. 2) and the formation of new strands until the nucleus 
finally contains a close network (Figs. 3 and 4) in the meshes of which lie 
one or two large nucleoli, and frequently several smaller bodies of a similar 
nature. While on the evidence from their history it is reasonable to suppose 
that the chromatic bodies are chromosome centres or prochromosomes, and 
probably maintain their individuality in the nuclear reticulum, nevertheless 
such individuality is not clearly evident in the resting nucleus of Oenothera. 
Synapsis. A contraction of the strands composing the nuclear reti¬ 
culum away from the nuclear membrane (Figs. 5 and 6) indicates the 
approach of synapsis. This contraction, as in biennis and grandijlora , 
proceeds until almost all of the strands in the network are drawn together in 
a close mass generally at one side of the nucleus and near the nucleolus 
(Fig. 8), which is sometimes quite surrounded by the web of threads (Fig. 7). 
This contracted and confused mass of strands, the arrangement of which 
cannot be followed with precision, constitutes the synaptic knot, a sharply 
marked stage extending over a considerable period, probably of one or more 
days. Occasional free loops extend from the contracted mass into the nuclear 
cavity, and the study of these, together with such structure as is apparent 
in the less dense regions, makes it clear that the synaptic knot is a close 
association of looped threads frequently anastomosing so as to form what 
appears to be a tangled mass. 
In such a complex association of strands it is not difficult to find threads 
that run closely parallel to one another for greater or less distances, or that 
are united at points. However, in a structure so involved and with elements 
so minute as are these threads, it is not safe to assume that these relations 
signify anything more than such a close association of strands as would 
naturally result from the contraction of the complexly looped and anasto¬ 
mosing thread or system of threads that constitute the nuclear reticulum. 
The evidence is entirely inadequate to support the interpretation that such 
parallel relations between the strands indicate the presence of two indepen¬ 
dent systems of threads (maternal and paternal spiremes), such as might be 
assumed to be associated with one another side by side. On the other hand 
it is impossible to say whether or not the coiled and twisted loops of the 
synaptic knot are parts of a single continuous thread or of a system of 
threads forming a reticulum ; if the former, it must be a thread of great 
length. 
As the process of synapsis proceeds it becomes evident that the 
threads thicken (compare Fig. 9 with Figs. 6 and 7), and from the later 
developments there is good reason for believing that this thickening is 
associated with a very material shortening of the thread system. The 
increased thickening of the strands, although clearly shown at the stages 
