4 BOTANICAL GAZETTE [JANUARY 
at certain parts, possibly by translocation of material from other 
parts. This gradual translocation of material tends to produce, 
out of the reticulum, smoother and smoother threads of somewhat 
uniform thickness, extending for some distance without branching. 
Such smooth strands are formed here and there from different 
parts of the reticulum. Of course, for some time these strands bear 
fine fibrils by which the various strands are connected into a single 
nuclear network (jig. 4). But finally these fine fibrils, which consist 
of chromatin, become disconnected; evidently the material is drawn 
into the strands, which naturally grow thicker on this account. 
The strands represent an early stage of the somatic chromosomes. 
When the chromosomes are just organized as a number of independ- 
ent elements (jig. 5), they are slender and very much curved, evi- 
dently lying in the position where they had first arisen out of the 
network as smooth thick strands. 
Owing chiefly to the curved nature of the chromosomes at their 
first appearance, it is difficult to prove that the place where a chro- 
mosome first appears is identical with the limit of the chromosome 
when last distinguishable in the telophase of the previous mitosis. 
This does not prove, however, that a chromosome may not appear 
in prophase in the same position in which it was last seen in the 
preceding telophase. 
LONGITUDINAL SPLITTING OF SOMATIC CHROMOSOMES 
The chromosomes thus formed are homogeneous and are strictly 
univalent during the prophase. The longitudinal splitting is first 
indicated very late in the prophase, just before arrangement at the 
equatorial plate. The process of longitudinal splitting is gradual 
and slow. Each chromosome, which has been compact throughout, 
becomes rather faintly stained in the central region, where the density 
of the aggregated chromosomes becomes less, although no change 
has taken place in the contour (fig. 6).. Then contractions occur 
simultaneously along the two lateral lines on opposite sides of the 
strand, where the structure has already become looser (fig. 7). The 
constriction proceeds inward from the two opposite lateral lines and 
meets in the center, thus dividing the chromosomes longitudinally 
into two similar halves. The longitudinal halves of each of these 
teas 
Rey Spee a: 
