560 Mor?rrieR AND NOTHNAGEL: CHROMOSOMES OF ALLIUM 
Later they separate and become irregularly distributed within 
the nuclear cavity (Fic. 14). At the same time they show a 
tendency to untwist, and as this is brought about the fact that 
many represent loops of the spirem is strikingly manifested. 
Bonnevie figures the looped and twisted condition of the 
bivalents, but the manner in which they originate from the spirem 
is not satisfactorily shown. In fact Bonnevie does not seem to have 
taken cognizance of the stages which we have described as the 
rearrangement of the spirem, or the second contraction, and this 
author has not, in our opinion, shown how her FIG. 33 is derived 
from FIG. 30. In our opinion her figures not only disprove the 
very thing she attempts to demonstrate, but lend support to the 
view set forth in the foregoing paragraphs, namely, that the two 
members of each bivalent represent pieces of the spirem that were 
previously arranged end to end and not the longitudinal halves 
of parts of the spirem. 
At the time of cross segmentation of the spirem the chromo- 
somes have attained their largest size.* During the formation 
of the spindle and later they seem to undergo a condensation by 
which their size is much reduced. Their number is seven or 
eight. While seven only were counted in some cases, the writers 
are inclined to regard eight as the correct haploid number. 
FROM SPINDLE TO DAUGHTER NUCLEI 
As is so well known, the spindle develops first as a multipolar 
complex which gradually becomes bipolar. Within the complex 
of spindle fibers the chromosomes are usually crowded so that the 
transition from the tightly twisted state of the two members to 
the large open ring-shaped structure that appears rather constantly 
in the equatorial plate of the mature spindle, is not readily followed 
(Fic. 15-17). The ring-shape of the bivalents in the spindle is 
doubtless the most striking phenomenon in the whole mitotic 
process of Allium cernuum. This form of chromosome is brought 
about by the fact that the chromosomes are attached to the spindle 
fibers at a point about midway between the ends. As the two 
members of each bivalent thus attached are drawn apart by the 
* The reader should bear in mind that Fic. 9, 1o, 11, 12, 14, 18 and 19 are more 
highly magnified than Fic. 13, 15, 16, 17, etc. 
