MOTTIER AND NOTHNAGEL: CHROMOSOMES OF ALLIUM 559 
radially loops and straight stretches of the cord, the latter with 
free ends. In Allium cernuum we do not find this typical appear- 
ance observed in Lilium. There is usually a tendency for the 
spirem to mass or become more closely entangled near the center 
of the nuclear cavity with a looping in the freer parts as shown in 
FIG. Io and 11, which represent the less complicated condition, 
but, as a rule, the entanglement is so complicated that it is not 
possible to follow definitely more than a few loops or turns of the 
eatire cord. If, for example, the spirem of FIG. 10 or II were 
bunched together more closely near the center with a greater 
twisting of the loops, we should have the more complicated state 
referred to above. The complexity of this step is increased by the 
fact that the spirem usually becomes more lumpy, or thicker places 
alternate with others more attenuated, just prior to, or as this 
rearrangement is ushered in (FIG. 9). Sometimes when the 
rearranged condition is not too confused, the spirem seems to be 
undergoing cross segmentation (FIG. 10), but whether this is the 
rule we are unable to say. It is certain, however, that in all, 
or in nearly all cases, the transverse segmentation of the spirem is 
accomplished during the entangled condition, or the stage of the 
second contraction. As segmentation is taking place there is 
always a violent twisting about each other of the two members of 
the bivalents, for as soon as the bivalents can be recognized as such, 
they invariably present the appearance of FIG. 12, save that they 
are more closely bunched together. Ordinarily they are heaped 
up in a morecompact mass. For the illustration we have selected 
a nucleus in which a less entangled massing of the bivalents is 
present (FIG. 12). 
That each bivalent, or the majority of them, represents a 
loop of the spirem, the two sides of which have twisted about each 
other, and not the two halves of the longitudinally split spirem, 
is in our opinion beyond question. The longitudinal split of the 
thread seen in the spirem disappears from sight, and, if it be present 
during the stages described, the two halves are so closely applied 
that no trace of the fission can be seen. Almost without exception 
the two members of each bivalent are twisted about each other, 
some tightly, others loosely (F1G.12,13). Inall cases the bivalents, 
as soon as formed, are massed and entangled into a confused heap. 
