288 BOTANICAL GAZETTE [OCTOBER 
however, soon come together again, after which all traces of the 
split are usually lost, although sometimes evidences of it may be 
apparent even after the formation of the chromosomes. 
After the two halves of the spireme have come together, it 
begins to contract. This contraction continues until the spireme 
shortens very considerably (fig. 33). The spireme at this stage 
has the appearance of a continuous thread, the ends of which are 
probably free. The spireme finally segments into five somewhat 
elongated chromosomes (fig. 34). Each of these chromosomes is 
probably bivalent, since the nucleus received five chromosomes 
from each of the two nuclei which by fusing gave rise to it. The 
bivalent condition, however, is not indicated by the form of the 
chromosomes. In this they are probably similar to those of most 
plants. In Peperomia (Brown 4), however, the two halves appear 
during the heterotypic prophase, as separate chromosomes con- 
nected by linin fibers; while in Oenothera (GATES 17) the diploid 
number of chromosomes appears at the same stage, and in this case 
some of the chromosomes may not be arranged in pairs. 
As the spireme contracts, linin fibers appear within the nucleus 
(fig. 33). Along those fibers, and especially in the early stages, 
there are small granules which have the appearance of chromatin. 
They usually stain less densely than the chromatin of the spireme, 
but frequently they are large and numerous enough to make the 
fibers along which they are scattered resemble the spireme. It 
was not possible to tell whether the substance of these granules 
passed to the chromosomes or took part in the formation of more 
linin fibers, but since as they disappear the number of linin fibers 
increases considerably, it seems probable that part of the granules 
take part in the formation of the fibers. No evidence of the forma- 
tion of these fibers from the linin of the spireme by the migration 
of the chromatin has been observed, but since the continuity of the 
spireme in the early stages is doubtful, and these fibers may resem- 
ble the spireme very closely, such a possibility, while not probable, 
can hardly be said to be excluded. It is certain, however, that 
most of these fibers which will later on take part in the formation 
of the spindle are formed de novo. 
As the spindle fibers increase in number, they become connected 
