518 THE CYTOLOGY OF THE SUGARCANE 
The consequence of chromosomes remaining univalent is — as was 
pointed out already — that the nuclei of the dyads are formed by 
numbers of chromosomes far superior to the haploid number. 
In the homotype division these abnormally high chromosomenum- 
bers are not reduced to the normal number. It appears that the chro- 
mosomes, arisen in the heterotype division by fission of univalent chro- 
mosomes, are in part distributed in the homotype division over the 
poles, according to chance, while the rest splits longitudinally. More- 
over sometimes not all chromosomes are taken in by the tetrad-nuclei. 
In these nuclei we will therefore, in exceptional cases only, meet with 
numbers of chromosomes corresponding to the normal haploid number. 
It is moreover almost unthinkable that the combination of chromoso- 
mes entering them should correspond to the combination present 
in a normal set of chromosomes. The chance is infinitely greater, that 
some chromosomes, essential for the normal development of the pollen- 
grain or for the fertility of the sperm nucleus, will lack in such sets, 
while other chromosomes will be overcomplete. Speaking theoretically 
the chance is great, that after such an abnormal reduction-division, 
none or but few fertile pollengrains will be formed. 
The investigation of EK 28 showed moreover, that when very many 
univalent chromosomes are present in the pollen-mothercells, complete 
or incomplete fusion of the dyad-nuclei can take place. This results in 
the formation of dyadcells without nuclei; of such with nuclei which 
contain a number of chromosomes far inferior to the haploid number; 
of others with abnormally high chromosome-numbers, which fre- 
quently surpass even the diploid number. Subsequently there arise te- 
tradcells without nuclei, tetradcells with a very small number of chro- 
mosomes and such with very many chromosomes in entirely abnor- 
mal combination. It is hardly necessary to state that the chance of the 
formation of normal nuclei is very small under these circumstances. 
Three of the five EK 28-inflorescences, which I had chosen in 1923 
for cytological investigation, came from a plot of cane which had 
grown well. The inflorescence which so frequently showed nuclear fu- 
sion in the dyads also came from this lot. From the same plot 8 inflo- 
rescences were selfed — for future hybridisation-experiments — 5 of 
them gave no seedlings at all, one gave two seedlings and 2 one seedling 
each. This of course points towards an extremely poor fertility of these 
inflorescences. Yet it is known, that EK 28 sometimes gives many 
4 
