YUZO HOSHINO 265 
the secondary determiner or determiners. The difference of variation types 
between D and "E, in Table 13 (see page 251), and P and Q in Table 17 (page 
255) can be understood, if we assume that the grand-parents of D ^'uid P 
become homozygous in the earlier secondary determiner or determiners, while 
those of E and Q become homozygous in the late secondary determiner or 
determiners. But the variation types of M, N and 0 hi Table 15 are not to 
be explained by a simple assumption of the presence of a secondary determin- 
er or determiners, as each of the above three raisings has its own proper 
variation types in earlier constant families, while the variation types of the 
later constant families are uniform in all three raisings. We hope, by further 
experiments, to be able to cast a clearer light on this contamination hy- 
pothesis. 
On Transgressive Inheritance. 
Throughout our experiments with peas and also with rice, we did not 
find a single case of transgressive inheritance. On the proposed two-factor 
hypothesis, however, we might expect the occurrence of such an inheritance. 
In the crosses between the late and early constant, or between either of these 
two constants and either of two intermediate constants (the early intermediate 
and the late intermediate), we can not expect any transgressive inheritance, 
but when we cross two intermediate constants (aaBB and AAbb) we should 
expect transgressive inheritance to take place. To state it in general terms, 
transgressive inheritance occiu's oftener in the case of crosses between the 
varieties with narrower difference of character than in the case of crosses 
between those with wider difference. From the address of Nilsson-Ehle at 
the Fourth International Conference of Genetics held in Paris, 1911, we find, 
although he did not state it in so many words, that among his several cross 
breeds in wheat and oats, those breeds which were the descendants of the 
crosses between the varieties with comparatively wider difference in their 
precocity did not show transgressive variation, while those descended from 
the crosses between the varieties with narrower difference did. Even a cross 
