AND SPECIES-HYBRIDS OF THE GENUS SACCHARUM 285 
In the counts of diakinesis-nuclei the number varied from 46 to 51, 
in metaphase similar results were obtained. This would lead one to 
suppose the right haploid number to be 48, which would form a series 
with the numbers 40 and 56. There is however no question of 48 gemini 
being present in the PMC. The varying numbers are caused by the 
presence of unpaired chromosomes. It is difficult to count the chromo- 
somes in the diakinesis-nuclei of Chunnee. It is not the nucleolus which 
causes the difficulty, as this body is generally fully decolorised and 
transparent. The cause of the difficulty lies in the fact that frequently 
masses of chromosomes are not lying adpressed to the nuclear wall but 
in the nuclear cavity. In Saccharum officinarum the univalent chromo- 
somes are easily recognised in diakinesis because they are arranged in 
sets of two, this is not so in Chunnee-cane or at least not so clear. 
Fig. 73 shows a nucleus in diakinesis in which a large number of the 
chromosome-elements have clearly the shape of gemini. Of one ele- 
ment, visible at a low adjustement, the two chromosomes composing 
it, looked as if they had undergone longitudinal fission, which is not 
impossible as among other plants longitudinal splitting of chromosomes 
during diakinesis, to be completed in the homotype division, has occa- 
sionally been observed. 

Fig. 74. Fig. 75. 
Fig. 74 and 75. Pollenmothercells of Chunnee in diakinesis x 2300. 

