634 'Wisconsin Academy of Sciences, Arts,- and Letters. 
out from the polar region to the periphery of the nucleus oppo¬ 
site. They are united with one another at the pole where they 
are also attached to the nuclear membrane. The other ex¬ 
tremities of the loops are also in contact with and attached to 
the nuclear membrane. The loops being in this manner at¬ 
tached to the nuclear membrane at both extremities, when con¬ 
traction of the spirem takes place, the state of “tension nu- 
cleaire” occurs and is relieved when the spirem breaks which, 
as Janssens believes takes place at the weakest place, the point 
of juncture of the chromosomes. 
While a certain tension may exist in these nuclei the 
lack of polar organization makes a close comparison between 
them and Janssen’s figures out of the question. Further 
the method of transverse segmentation here cannot differ greatly 
from such segmentation in typical mitoses where no such ma¬ 
chinery seems necessary to the formation of chromosomes. 
As the spirem gets shorter and thicker and the central tangle 
at the same time becomes more dense, the longitudinal segmen¬ 
tation of the spirem becomes more and more distinct, until at 
the time of the maximum density of the central knot the spirem 
is clearly split throughout its entire length so that in many 
places the halves diverge from one another for short distances. 
It is at this period that the spirem segments transversely in the 
peripheral region to form the heterotypic chromosomes. I did 
not find any instances in which the transverse segmentation 
took place elsewhere than at or near the periphery. In no case 
was I able to find free ends of the spirem in the central region 
of uncut nuclei. 
Immediately following transverse segmentation the longi¬ 
tudinal halves of the spirem are seen to be twisted around one 
another forming the familiar strepsinema condition (Fig. 21). 
Previous to segmentation the twisting seems to be obscure or 
lacking. It is probably due to unequal shortening of the outer 
and inner parts of the halves of the split spirem. When the 
spirem is continuous or nearly so there is little chance for 
twisting to take place but transverse segmentation makes it pos¬ 
sible. 
