THE SPERMATOGENESIS OF LEP1DOSIREN PAKADOXA. 15 
that in which the attachment of the chromosomes to the 
nuclear membrane first gave way. Whether this clumping 
together occurred during life or whether it needed the 
additional strain of the fixation to bring it about is probably 
an unimportant matter. 
The fact that the onset of synizesis coincides with the 
entrance of the nucleus iuto strepsinema causes the details of 
the latter to be less diagrammatic than in many other forms, 
e.g. Tomopteris, Salamandra (Schreiner), Dicro- 
coeliutn (Dingier). The separation of the conjugants also 
takes place very quickly along their whole length, except 
(for a time) at their ends, so that the pachytene threads are 
rapidly converted into very long rings. This is shown in 
PI. 2, fig. 16, which is a polar view of a strepsitene nucleus. 
Instead of the free ends of the pachytene threads we find that 
each of them has split and forms a long ring, the greater part 
of which is hidden in the synizetic mass below. In the case 
of most of the threads the conjugants have separated com- 
pletely, so far as can be seen, except at their ends, but some 
of them can be seen in the act of separating, being' still fused 
together in places. The shortening and thickening of the 
threads has, of course, been proceeding steadily since the 
first onset of the prophase, so that the separating conjugants 
are far thicker than they were at the time when they fused 
in zygonema. The chromosomes have lost the beaded appear- 
ance they possessed in the bouquet stage. 
From the study of the preceding and following stages as 
well as by aualogy with other forms we know that the rings 
are present in the reduced number. In the stages imme- 
diately succeeding strepsinema in Lepidosiren a very 
curious process takes place. The rings break through first at 
one joint (i.e. point of attachment of conjugants) to form 
open loops, and then at the other to form perfectly separate 
chromosomes. In late diakiuesis we thus get the somatic 
number of chromosomes present, which show no signs what- 
ever of having been previously united. 
Figs. 17-24 illustrate these stages. In the interval covered 
