Ikeno. — On Hybridization of some Species of Salix. II. 183 
of segregates was somewhat different from what we might have expected in 
the case of usual Mendelian segregation, I could not then decide definitely 
whether we have here really to deal with such or not. In 1918 Heribert- 
Nilsson 1 published an elaborate memoir containing the results of his 
extensive experiments on Salix hybrids ; he was able not only to prove 
the segregation of various characters, but also to discover the usual 
Mendelian ratio in many cases. My experiments concerning the catkin 
characters above mentioned have shown that their segregation occurs 
according to the simplest Mendelian ratio 3 : 1, though more or less obscured 
by the imperfection of dominance. Accordingly these experiments are 
a further contribution towards establishing the fact that the segregation of 
the various characters in Salix hybrids follows Mendel’s law. 
III. Formation of Parental Form. 
The hybridization multinervis xgracilistyla done in 1911 has given 
rise to almost fifty offspring, all of which were, contrary to our expectation, 
not hybrids, but female multi tiervis . 2 
I have myself had no doubt about the actual production of multinervis 
progeny in this case. The fact was, however, so unexpected that 
others may reasonably be in doubt about its reality. So I repeated 
the same hybridization in 1918 and 1919. The results of hybridization in 
1918 are contained in Table I (see p. 176). Though, as is usually the case 
in this hybridization, each catkin has produced only few seeds — often 
none — yet I have obtained a comparatively large number of progeny — more 
than ninety plants in all ; but the number of multinervis was small — in fact 
I had only two (Table I, Nos. 26 and 29). On the contrary, the hybridiza- 
tion carried on in 1919 has yielded very few seedlings, but almost all were 
multinervis progeny, thus : 
Table VII. 
No. of the Female Parent * r - „ . 
{multinervis o). N °- °f Catkins. 
No. of Progeny 
(; multinervis ). 
I. 
35 
2 
36 
1 
41 
1 
II. 
4 
1 
. 7 
4 
Total 
9 
Gracilistyla used as the pollen plant in the above experiments was the 
same in both cases, and identical with what we have employed in the 
experiments whose results are shown in Table I. 
No progeny contained in the above table bear flowers yet, but that 
all of them are multinervis is quite unmistakable on account of their 
characteristic leaves, &c. Besides them, some plants which are not multi- 
1 Lunds Universitets Arsskrift, N.F., Afd. 2, Bd. xiv, No. 28, 1918. 2 1 c>} p> gI- 
P 
