YUZO IIOSHINO 
249 
there are only 6 families which vary like the above mentioned families 
(No. 32, 33, 37, 39 and 41 in Table 9, and No. 33 in Table 11). It is 
quite probable that the parent plants of these 6 families belonged to 
the early flowering group geneticall}', but for some cause their flowering 
time had been so retarded as to make them fall into the late flowering 
group. 
3. The variation ranges of all other variable families descended from the 
late-flowered F2 parents reach the variation range of G- P- Some vary 
from the range of M- P- to that of G. P-; as in the ca.se of the F.^ fami- 
lies ; some vary within a combined range of the early intermediate con- 
stant and G. p.; while some show such a narrow variation range as to 
begin from that of the late intermediate constant. 
4. There are a few exceptional families descended from the late-flowered F.^ 
individuals, whose variation ranges do not reach that of G- P- but only 
that of the pseudo-late constant families (No. 60, 62, 67, 71 and 77 in 
Table 9, No. 34, 35, 47, 56 and 57 in Table 1 1). 
5. Comparing the variation types of all variable families in Table 9 with 
those in Table 10 and Table 1 i, we did not find any difference in charac- 
ter between them. So, again, we can assume that the hereditary differ- 
ence of the pure line character is qualitative and not quantitative. 
Raisings. 
In 1913. 
This year, as it needed a comparatively large area for raising the F3, the 
results of which had already been shown in Table i i (Pis. XV and XVI), and 
as the available space was limited, only the F_^ progenies of the following F3 
families in Table 10 (PI. XIV) were grown, as preliminary raisings: — 
No. 8 (I X [.4-3 ) (A) 
No. 18 (I X 1.4.4 ) (B) 
No. 45 (Ix i.4.14) (C) 
No. 43 (I X 1.4.9 ) (D) 
No. 50 (I X 1.4.13) (E) 
No. 38 (I X 1.4.2 ) (F) 
