YUZO HOSHINO 
the occurrence of such deviation does not militate against proposed two-factor 
hypothesis. 
We shall next examine the experimental results in F^^ raisings. The 
variation types of C, D, E and F in Table 13, and M, N and 0 in Table 16 
can be interpreted if we assume that the F2 grand-parents were aABB in their 
zygotic constitution, taking gametic contamination as an admitted fact (pages 
251 and 254). In the same way, we can assume B in Table 12 and G and H 
in Table 14 to be the variation types of the offsprings of aabB (pages 250 
and 252); I in Table 14, of aAbb (page 252) ; P, Q, and B. in Table 17, of 
AAbB (page 254) ; and S, T and U in Table 18, of aAbB (page 255). A in 
Table 12 and K in Table 15 indicate the homozygosity of their grand- 
parents (the early intermediate constant aaBB.)- The variation types J and 
L in Table 1 5 might be interpreted as those of the progenies of the homozy- 
gous aaBB, if we assume that their F^ grand-parents by gametic contamina- 
tion produced such offsprings as vary within a small range. 
Gametic Contamination. 
It has been affirmed by the pedigree raisings of the parent varieties from 
/910 to 191 3, that the character of the flowering time was quite fixed and 
unchangeable in the parent varieties. (Tables 4, 5, 6) From this fixedness 
of the flowering character it should be expected, according to the common 
Mendelian interpretation, that all families descended from those individuals 
which had the same zygotic constitution, would produce the same variation 
type. But our actual experimental results do not fulfil this expectation, as 
long as we keep the proposed two-factor hypothesis. 
In the F3 raisings, there has been frequent occurrence of the pseudo- 
early and pseudo-late constants. Such is not admissible on the two-factor 
hypothesis when gametic purity is retained. We took B in Table 12, as 
variation types of the segregation of aabB. but the variation means of all six 
earlier constant families are larger than that of M- P- The variation types of 
four raisings, C, D, E and F in Table 13 and of three raisings, M, N and 0 in 
