


THE CYTOLOGY OF THE SUGARCANE 509 
The irregular division was especially apparent in the anaphase, 
because a large number of univalent chromosomes lagged behind on 
the spindle. Fig. 15 pictures this. 
At the poles two small groups of 
chromosomes are seen, almost in 
telophase; already the outer rows 
only are still dictinctly à part. 
Between these two groups a large 
number of chromosomes are 
scattered over the whole spindle. 
The chromosomes, visible at a 
very low focus, have not been 
drawn, so as not to make the 
picture too complicated. The 
chromosomes at the poles have Fie. 15. 
been derived from gemini, those Anaphase of the heterotype division of 
EK 28, viewed from the side. x 2300. 

on the spindle are all univalent, 
part of them has already split up into much smaller ones, some of them 
show the beginning of splitting. 
The, as yet, unsplit univalent chromosomes correspond in size with 
the chromosomes derived from gemini. 
Anaphases as the one here pictured, formed the rule in these prepara- 
tions. A similar division is a reduction-division, only in regard to a rather 
small number of chromosomes, a somatic division in regard to the lar- 
gest number of them, in which halves of unpaired chromosomes move 
towards the poles. In the case of a real reduction-division the chromo- 
somes reach the poles in haploid number, in a case like this, in a num- 
ber closer to the diploid than to the haploid one. 
In homotype divisions also very great irregularities were observed. 
Exact counts were not possible, but it was very clear that many more 
than 40 chromosomes were present in the nuclear plates. Halves of uni- 
valent chromosomes, besides chromosomes derived from gemini, must 
occur in the nuclear plates. One might now assume, that the halves of 
‘those chromosomes, which had not paired in the heterotype division, 
had subsequently formed pairs in the homotype division. In that case 
the chromosome-reduction, would, in regard to a large part of the 
chromosomes, have been passed from the heterotype division to the 
homotype one. Such however is not the case, for, if it were so, 40 chro- 
