178 Ikeno . — On Hybridization of some Species of Salix. II. 
actually the case, inasmuch as I have got by this crossing not only 
G-type, but a certain number of M-types (see Table V, Expt. 6 , below). 
Such a fact, as well as certain other reasons which would not be needed 
here to be enumerated, have led me to abandon my hypothesis, at least 
in the form in which it was expressed in my former paper. How can 
we then explain the phenomenon ? For the reasons which I will not 
repeat here I was formerly unwilling to accept the ‘ imperfection of domi- 
nance * for its explanation. 1 But if the view of ‘ invisible factors ’ is to be 
abandoned, I think that there will be no better means than to resort to that 
of the ‘ imperfection of dominance ’, for the latter will explain most simply 
what we observe in respect to Salix hybrids ; though I am still of the 
opinion that the interference of certain other factors, either visible or 
invisible and variously combined, may be the ultimate cause of this 
phenomenon. What is discussed below is consequently founded on the 
view of the ‘ imperfection of dominance ’. 
We have seen in Table I that in F 1 of multinervis x gracilistyla we 
have 83-3 per cent. G-types and 16-7 per cent. M-types. We may regard 
the densely hairy condition of the catkin of gracilistyla to be dominant to 
the less hairy condition of that of multinervis as a rule, so that if we repre- 
sent the former by D and the latter by R, all F 1 hybrids will agree in being 
DR genetically. The dominance here is not absolute ; consequently only in 
the 83-3 per cent, is the dominant character observed, whilst in the remain- 
ing 16-7 per cent, it fails, and lets the recessive character appear externally. 
In short, the G-type as well as the M-type progeny in F 1 are genotypically 
equal, though phaenotypically different ; we have evidently no segregation in 
this generation. Moreover, Table I shows us that many numbers therein, 
each of which corresponds to one catkin, contain both G- and M-types. 
Hence we may conclude that not only one plant individual, but also even 
one and the same catkin, possesses at the same time both gametes where 
usual dominance of D over R prevails and those where it fails. 
2. F 2 Generation. 
In order to obtain the F 2 progeny I have carried on the three follow- 
ing crosses between the G-type and M-type progeny. 2 The results are 
shown in the following table 3 : 
1 1. c., p. 48. 
2 Only in Expt. 3 is the male parent derived from F 2 (1. c., p. 42). 
3 The number of progeny in Expts. 1 and 2 is the same as in my former paper (1. c., pp. 42-3), 
but that in Expt. 3 is a little larger, because I have added afterwards those plants which have 
borne flowers in 1919 and 1920. In calculating the percentage the new type was not taken into 
account. 
