AND SPECIES-HYBRIDS WITHIN THE GENUS SACCHARUM 317 
Primula sinensis has in the somatic cells 24 chromosomes, in the ga-’ 
metes 12; Primula sinensis gigas has in the somatic cells 48 and in the 
gametes 24 chromosomes. The tetraploid Primula sinensis gigas proved 
to be less fertile than the diploid P. sinensis, a sufficient number of 
good seeds, to keep this race alive, was however formed. 
The tetraploid Primula-hybrids, which originated from diploid Pri- 
mula-species, possess a normal reduction division, in the prophase ge- 
mini only and no unpaired chromosomes occur ; they arose from parents 
with equal diploid chromosomenumbers. 
Compared with the tetraploid hybrids and with the hybrids between 
Saccharum officinarum and S. spontaneum hetero-diploid hybrids gene- 
rally are less fertile. In the prophase of the division of the gonotokonts 
of the former, all or nearly all chromosomes proceed to a complete pai- 
ring. Among the hetero-diploid hybrids there is never question of a 
pairing of all chromosomes. Sometimes no gemini whatever are even 
formed by them, as in the case of the Digitalis-hybrids or a very limi- 
ted gemini-formation only takes place, as in the case of Polypodium 
Schneideri. In such hybrids the division of the gonotokonts proceeds in 
an entirely irregular manner, the chromosomes are distributed without 
any regularity, merely according to chance, over the poles. No wonder 
therefore, that these hybrids are sterile. 
Among hybrids, in which the division of the gonotokonts proceeds 
according to the Drosera-scheme, the unpaired chromosomes can behave 
very differently. In the case of Drosera and Oenothera they are distri- 
buted in the heterotype division, without any splitting, over the poles 
and may form dwarfnuclei ; in the homotype division, they split longi- 
tudinally and contribute sometimes to the formation of the tetrad- 
nuclei, while, in other cases, they form dwarf-nuclei and are thus elimi- 
nated; in the case of Oenothera they may fragment moreover into seg- 
ments both in the anaphase of the heterotype and in that of the homo- 
type divicion. Among Hieracium unpaired chromosomes may split al- 
ready longitudinally in the heterotype division and this splitting may 
or may not be repeated in the homotype division. As a consequence 
these hybrids can form tetradnuclei with very different chromosome- 
numbers. It does not appear impossible to me, that this may result in 
combinations of chromosomes which exclude viability on the part of 
the gametophyte, as might happen in the case of the Saccharum offi- 
