July, 1922] 
CLELAND — OENOTHERA FRANCISCANA 
401 
circle of four after a time (fig. 28). The circle of four, however, does 
not break up in diakinesis, but remains until the disappearance of the 
nuclear membrane, and even later, being often seen during the formation 
of the equatorial plate and even occasionally until metaphase (fig. 29). 
In some cases one or more of the rings remain linked to the circle of four 
until the end of diakinesis, but more usually we find in late diakinesis 
that they have all separated, so that we have the circle of four, and five 
pairs of chromosomes scattered about in the nucleus. One of the most 
interesting features of this stage is its remarkable clarity. The chromo- 
somes are clearly distinct from one another, often attached together only 
by narrow threads so that they stand quite apart. The presence of such 
a circle of four chromosomes has been observed before. Miss Digby (1912) 
has described this condition in one of the Primula hybrids. Her figures 
undoubtedly testify to its existence in her material, but it evidently does 
not show as clearly as it does in Oenothera franciscana. 
During diakinesis the nucleolus is plastered against the nuclear mem- 
brane and is fast disappearing. At times it appears vacuolated, at others 
it is evenly translucent and apparently empty except for the black-staining 
endonucleolus which is often still prominent. The nucleolus seems to be 
melting. The edges are becoming hazy and ill-defined, and the structure 
looks as though it were being dissolved away. I have seen the endo- 
nucleolus in a few cases apparently in the act of being liberated, through 
the melting away of the nucleolus, and it is very probable that this little 
body maintains its identity for some time after the nucleolus has dis- 
appeared. 
Metaphase 
It is not until the nuclear membrane is on the point of breaking down 
that signs of spindle formation become evident. The first indication of 
this formation is a sudden increase in density of the cytoplasm close to 
the nuclear membrane. At first no fibrils are to be seen. These do not 
appear until the nuclear membrane begins to dissolve. Then they sud- 
denly make their appearance on all sides, and, as the membrane melts, 
penetrate into the nuclear region. Within a short time, a very prominent 
multipolar spindle is formed, with several thick clusters of prominent 
fibrils. As these pass into the nucleus, and in and around the chromo- 
somes, those of each cluster anastomose in various ways with fibrils from 
other bundles (fig. 29). 
While the spindle is being formed and the nuclear membrane is dis- 
appearing, the chromosomes rapidly shrink, until they are less than half 
the size which they have presented during diakinesis. They have probably 
had a somewhat spongy texture, and are now being greatly compacted. 
During the whole of the period when the spindle is still multipolar, the 
circle of four remains as a rule unbroken, and even shows frequently one 
or more rings still Hnked to it. During this stage also a small but prominent 
