Ikeno.—On Hybridization of some Species of Salix. II. 187 
In T.halictrum pnrpnrascens x as well as some species of Hieraciutn 
(■ pilosella , excellens , &c.) 2 it is well known that there are two kinds of 
ovules, viz. those which require fertilization for seed formation, and those 
which do not. It will naturally be quite the same in Salix multinervis . 
As we have here to deal in all probability with somatic parthenogenesis 
in the sense of Winkler, 3 it follows that the former kind of ovules, which 
have undergone the reducing division of chromosomes during their de- 
velopment, become hybrids by pollination with foreign pollen, while the 
latter kind, which have undergone no such process, develop to multinervis 
without being fertilized. Moreover, that these two kinds of ovules are 
contained at the same time not only in one plant individual, but also in one 
and the same catkin, is clearly seen from the examination of No. 26 in 
Table I (see p. 176), inasmuch as in this number the progeny which are 
derived from one single catkin contain, besides one multinervis, four G-types 
(i. e. hybrids). 
In my former paper I was led to the conclusion that in our case 
we have to deal with pseudogamy, i. e. parthenogenesis induced by the 
stimulus of foreign pollen, gracilistyla in the present case. This was the 
natural consequence deduced from the results of my former experiments, 
because then any catkin covered with a bag was found not to be able to bear 
a single seed. Since, however, our new experiments have shown us that in 
certain cases, though extremely rare, catkins wholly prevented from any 
pollination are able to produce a few seeds, my former conclusion is of course 
not quite right. The following remarks may, however, be made in conse- 
quence of the results of our new experiments. Although very rarely, the 
apomictic development without any pollination is possible, yet we have seen 
that the number of seeds then produced was so scanty that pseudogamy 
might also be not impossible in certain cases. For instance, in our experi- 
ment done in 1911, where the pollination with gracilistyla was several times 
carried on, I have got almost fifty midtinervis individuals from a few 
catkins, while in the experiments done in 1920 and 1931, when no such 
pollination was practised, only five apomictic seeds which were able to 
germinate were obtained out of more than 700,000 + 370,000= 1,070,000 
ovules! 4 It is not improbable that although few ovules may develop 
parthenogenetically without the stimulus of foreign pollen (autono- 
mous parthenogenesis), yet ovules may in many cases be induced to 
1 Overton : Bot. Gaz., vol. xxiii, 1902, pp. 363-75 ; Ber. d. Deutsch. Bot. Ges., Bd. xxii, 1904, 
pp. 274-S3. 
2 Ostenfeld : Botanisk Tidsskrift, Bd. 27, 1906, pp. 225-48. Rosenberg: ibid., Bd. xxvii, 1907, 
pp. 143-70. 
3 1. c., p. 17. 
4 In the experiment carried on in 1919, when the pollination by gracilistyla was practised, 
I got only nine multinervis (see Table VII), but then only a very few catkins were used for experi- 
ment, and it is hardly doubtful that had I experimented upon a larger number of catkins I should have 
been able to obtain a much larger multinervis progeny. 
