542 
K. IIUGGLKS GATES ANJJ NESTA THOMAS. 
the same size as the other chromosome-halves. Certain other 
cases which show this even more conclusively will be 
described later. In fig. 25 the cell, nuclei and chromosomes 
are all distinctly smaller than in fig. 24^ and the chromosomes 
show very little evidence of a split. The left-hand group 
contains 7| chromosomes, while the right-hand group con- 
sists of 6 whole chromosomes, a half- chromosome and a 
small fragment. There is in addition a chromosome in the 
cytoplasm near this group, which has been left out of the 
nucleus in interkinesis. This makes the full quota of 15 
chromosomes. The origin of the fragment is obscure, but 
the figure shows that even fragments of chromosomes may be 
distributed occasionally to the daughter-nuclei. Unless 
afterwards extruded from the pollen nuclei, such fragments 
would no doubt affect the later development of the indi- 
vidual; they might remain independent or become attached 
to or fused with one of the other chromosomes. These 
unequal divisions of chromosomes furnish an obvious source 
of variability within certain limits — such variability as is 
found in the lata-se m i lata series of forms. 
Fig. 26 is a profile view of a homotypic spindle in very 
early anaphase when the chromosomes are just separating. 
In figs. 27-31 are represented later stages of the homotypic 
anaphase, showing the distribution of chromosomes to the 
pollen tetrads. Fig. 27 a and h shows the two homotypic 
spindles in a mother-cell, degenerating chromosomes being left 
behind in the median region of both spindles. In h there are 
8 chromosomes in a group near each pole, and in addition 
two fragments which stain less deeply and are evidently 
degenerating without being drawn to the pole. In a there 
are only 6 chromosomes at either pole, and again two frag- 
ments remain behind to degenerate. It is very probable that 
these fragments are derived in each case from a half of one 
chromosome which underwent fission on the heterotypic 
spindle. There is also much variation in the size of the 
chromosomes in this cell, those in h being on the average 
somewhat larger than those in a. 
