SOMATIC CHROMOSOMES IN TRADESCANTIA 
347 
heaviest portions a few small vacuoles may for a time persist, but for con- 
siderable distances the thread is completely without them. 
The true split now develops in the slender threads. Almost as soon as 
a portion of a thread becomes sufficiently equalized a number of new 
openings appear, and it seems highly probable that they are the outgrowth 
of small vacuoles which are formed anew along the axis of the thread. 
Some of the openings elongate a little, making the thread clearly double for 
short distances (fig, 19); here for the first time a portion of a chromosome 
can be said to be truly double. It is a matter of extreme difficulty — indeed 
it is probably impossible — to tell certainly whether these small vacuoles and 
openings are all new prophasic developments or are in part retentions from 
the resting stage and hence from the preceding telophase. Soon after 
their formation the slender threads, so far as can be determined with the 
best optical equipment, are certainly single and devoid of vacuoles for long 
distances, whereas the vacuoles become very numerous later. After a 
comparison of many chromosomes in these stages the writer has concluded 
that in all probability a vacuole or open space is now and then retained 
from an earlier stage, but that the great majority of vacuoles and spaces 
which develop into a split in the slender thread are formed anew in the 
prophase. 
As the vacuoles increase in number and enlarge into openings extending 
through the thread, the latter shortens and thickens, and takes the form 
of an irregular ladder-like structure (figs. 20, 21). A nucleus in this stage 
(fig. 22) has a superficial resemblance to one in the early prophase stages 
(fig. 15), and some writers have confused them, omitting the important 
stages which intervene. In figure 15 the chromosomes are in the form of 
irregular alveolar-reticulate cylinders, as shown by their cross section 
(fig. 16), whereas in figure 22 they exist as thread- or ribbon-like structures 
partially split by a series of median openings and are clearly double in cross 
section (see the free and cut ends in figures 21-23). 
As already stated, Tradescantia appears to be less favorable for a study 
of the later prophases than Vicia. The open spaces in the chromosomes do 
not run together to form a continuous split so early as in the latter plant. 
The material of some of the cross pieces connecting the sides of the partially 
split threads gradually flows to the two sides where it accumulates in the 
form of paired lumps simulating divided granules as fully described in the 
account of Vicia. Many of them, however, remain unchanged until a 
very late stage, so that even after the threads have become much shortened 
and thickened to form the conspicuous heavy spirem stages some of them 
may show the openings not yet developed into a complete split (figs. 22, 23). 
In other chromosomes the splitting process has gone further, making them 
almost completely double (at right in figs. 22 and 23). In nuclei of these 
stages it is not difficult to discover several free ends which are not due 
to the microtome knife, so that certainly at this time, and probably at 
