YUZO HOSHINO 
257 
tain the following 9 different zygotic series, 5 heterozygous and 4 homozy- 
gous. 
Homozygous Heterozygous 
I aabb i^, , ^ . 2 aabB 1 , ^ 
■Karly flowering group 'Early flowermg group 
I aaBB > " 2 aAbb 
4 aAbB 
Late flowering group 2 aABB 
I AAbb 
I AABB 
Late flowering group 
2 AAbB 
Among the 5 heterozygous series, aAbB has the same zygotic constituent as 
Fj, and its flowering time must incline towards that of the late parent rather 
than intermediate between those of both parents (page 233 and 242); aABB 
and AAbB, which have become partly homozygous on account of the intro- 
duction of a factor from the late parent, must flower later than aAbB ; aabB 
and aAbb, which have become partly homozygous by the introduction of one 
factor from the early parent, must flower earlier than aAbB- Two homozy- 
gous series, aabb and AABB, have the same zygotic constitution as the early 
and the late parent respectively. As we have seen in that the dominance 
of lateness was incomplete, the following assumption on the flowering time of 
tlie i-emaining two homozygous series, aaBB and AAbb, is quite probable. 
aaBB, which is homozygous in the earlier determiners from both parents, may 
flower earlier, inclining towards the early parent, while AAbb, which is homo- 
zygous in the later determiners from both parents, may flower later, inclining 
towards the late parent. 
In actual raisings of F^ families, we could always divide the individuals 
into two groups, the early flowering and the late flowering, by taking the 
minimum frequency class, which was at a nearly intermediate position be- 
tween the parent varieties, as a line of demarkation. Now in the proposed 
hypothesis, we assume that the early flowering group consists of four series, 
aabb, aaBB, aabB and aAbb, and the late flowering group, of the remaining 
five series, aAbB, aABB, AAbB, AAbb and AABB- The ratio of the 
number of individuals in the former four series to those in the latter five series 
