5 36 Lewis.— The Behaviour of the Chromosomes 
already quite well established that the nucleolus is concerned in some way 
with the elaboration of the chromatin. 
My study of the nuclei as they approach and recover from synapsis has 
convinced me that the difference of interpretation prevailing among cyto- 
logists at present in regard to the conjugation of the spirems is due to a 
confusion of the stages. Cardiff’s (’ 06 ), Figs. 8, 9, 10, 12, 43, 54, 63 are, 
judging from my own experience with mitotic division, drawn from post- 
synaptic nuclei, although they are much more diagrammatic than anything 
I have encountered in any cells yet studied. 
Formation of the Chromosomes and their Distribution 
to the Daughter Nuclei. 
The manner in which the chromosomes are formed from the spirem 
presents the greatest difficulty of any step in the entire process of mitosis 
in these genera. The origin of the spirem and its behaviour subsequent to 
synapsis seems to preclude the possibility that the chromosomes represent 
pieces of the spirem in which the longitudinal fission has again become 
evident, although such a view was formerly held by those investigators who 
believed in a double longitudinal fission. This view is also held by those 
who interpret the fission of the prophase as a conjugation of maternal and 
paternal spirems side by side. Miss Ferguson (’ 04 ) maintained that the 
spirem, which is formed as described above, segments into pieces which 
equal the number of somatic chromosomes typical of the species. These 
segments then unite in such a way as to give rise to a sort of reticulum, 
which finally segments again to form the bivalent chromosomes. Each 
bivalent consists therefore of two somatics which have become united 
before final cross-segmentation. The chromosomes are arranged in the 
post-synaptic nucleus end to end, but the pieces which form the parts of any 
one bivalent are not necessarily adjacent to each other until after the first 
transverse segmentation. This view harmonizes somewhat incompletely 
with that of Farmer and Moore, Mottier, Schafifner, Juel, and others. 
The spirem of Pinus and Thuja , as indicated above, becomes a regu- 
lar homogeneous band in which only the slightest evidence of the longitudinal 
split remains. It is wound evenly throughout the cavity and consequently 
thrown into a number of loops (Figs. 10, 11). That the number of loops 
corresponds to the number of bivalent chromosomes cannot be stated 
definitely, although there may be some such relation. The spirem thickens 
very considerably and signs of cross-segmentation become quite evident. 
In some nuclei the number of loops is now almost identical with the reduced 
number of chromosomes (Fig. 14). This figure is drawn from Thuja. The 
nucleus is shown almost entire and the loops are quite as regular as any yet 
observed for this or any other genus. It is also quite evident here that the 
sides of the loops are double in certain places, although usually the double 
