YUZü HOSHINO 247 
of M- P-j and No. 19, the late constant, equivalent to G. P- Other families 
are more or less different from those of the parent varieties, but none are 
transgressive, being all intermediate between the parent varieties. Among 
the.se intermediate constants, No. 4 is very near to the early constant, and 
Nos. 16 — 18 to the late constant. We shall designate such families as pseudo- 
early and pseudo-late constants respectively. Other remaining intermediate 
constants are to be classified into two groups. Nos. 5 — lo are to be grouped 
together as the earlv intermediate constant, and Nos. 11 — 15 as the late inter- 
mediate constant. And it will be seen that the F., parent individuals of all 
families of the former group belonged to the early flowering group and those 
of the latter group to the late flowering group. (Compare parent classes of 
these families and minimum frequency classes in Table 7) 
In Table 10 (PI. XIV), Nos. 1 and 2 are considered early constant and 
No, 3 pseudo-early constant. The 5 families, Nos. 12 — 16, have the varia- 
tion t\-pes very similar to that of G- P-, only 2 of them (No. 12 and No. 13) 
have a few individuals which flowered a little earlier than G- P- and may be 
disignated as the pseudo-late constant. All other intermediate constants are 
to be classed into two groups just as in the case above mentioned. The 5 
families, Nos. 4 — 8, all descending from the early flowered plants, are 
considered as belonging to the early intermediate constant, and the 3 fami- 
lies, Nos. 9 — II, descending from the late flowered F.^ plants, as belonging 
to the late intermediate group. Thus, the mode of distribution of constant 
families is quite similar in both Table 9 and Table 10, but in the variation 
types of constant families, there is some difference between them. The 
variation types of the early constant and the early intermediate constant 
families in both Tables do not differ much from each other, but the late inter- 
mediate constant and the late constant families in Table 9, where the late 
pure line of G- P- "^^'as one of the parents, take a distinctly later position 
than those in Table 10 where the early pure line of G- P- was one of the par- 
ents. This indicates very clearly the hereditary difference of the two pure 
lines of G- P- From the similarity of distribution of constant flimilies, how- 
