1 86 Ikeno . — On Hybridization of some Species of Salix. II. 
raged several times in March and April. As catkins on such branches 
must necessarily have been left out of consideration, the number of those 
available for our experiment was considerably diminished. Nevertheless, 
we had 1,261 catkins which should contain no less than 1,261x100x3 
= 378,300 ovules in all, i. e. more than half the number of ovules 
experimented upon in 1920. All these catkins were covered with 
bags for more than two months. As in the former year I saw not 
unfrequently ovaries which grew somewhat more intensely than others, 
but generally they either simply shrivelled up gradually, or opened 
and proved themselves to contain no seeds at all. Only on some 
catkins of the tree designated as No. 1 in Table VIII did I get eight 
seeds. Of these, three came to germination, and one began to germi- 
nate but soon ceased to make any further growth. All remaining seeds 
refused to germinate. 
In the two new experiments just mentioned I was thus able to get only 
very few seeds which have the power of germination, but nevertheless they 
have proven beyond all doubts that in our case the apomictic development 
of ovules without the application of any pollen is a matter of possibility, 
though it is of extraordinarily rare occurrence. 
In consequence of the results of the new experiments just described my 
former view concerning the development of multinervis after the pollination 
with gracilistyla, as stated in my former paper, must necessarily change. 
To explain the result of our new experiments there are two possibilities, 
viz. the embryo formation from nucellar cells and the parthenogenesis, each of 
which is either autonomous or induced by the stimulus of foreign pollen. 
The definite conclusion whether the apomictic development in question will 
occur according to the first or the second mode just enunciated would 
naturally be possible only after a comparative cytological examination 
of the development of normal as well as apomictic ovules. Since, as above 
stated, ovules of the latter class are of extremely rare occurrence, it would 
be necessary for the purpose to examine innumerable specimens made by 
sectioning an enormous number of catkins, and even then there would 
be no great chance of meeting with such ovules. Thus we see that in the 
present case the cytological examination must be almost an impossibility. 
The embryo formation from nucellar cells is generally accompanied by 
polyembryony, as in Funkia , Citrus , Opuntia , &c., &c. Since I have never 
met with this in Salix , although I have observed the germination of 
many thousand seeds of various species of Salix , either pure or hybrid, 
including S', multinervis , I am rather inclined to the view that as the 
cause of apomictic development in our case parthenogenesis is much more 
probable than embryony. All the following discussion is accordingly 
founded on the supposition that we have here to deal with partho- 
genesis, either autonomous or induced. 
