218 MENDELISM 
With three such factors we should expect only one 
of each parent type to appear among sixty-four 
individuals. 
There seems to be no reason for doubting that the 
total number of visible factors, even, which go to 
make up the total shape of a flower may be consider- 
ably greater than this; and the number of invisible 
allelomorphs upon which these depend may be pro- 
portionately more numerous still. In such a case, in 
an experiment where the total number of offspring 
grown was limited, types like the original parents 
might never be seen, simply because the sample taken 
was not sufficiently large. 
Again, we have to consider the possibility of a 
differential fertility among the various allelomorphic 
combinations. This might lead to an intermediate 
form, actually breeding true in spite of the fact that 
segregation was going on in its germ-cells. The 
possibility has to be borne in mind that among the 
offspring of two widely different species those indi- 
viduals which more nearly resemble their immediate 
hybrid parent may have a better chance of survival 
than the forms which have more characters in common 
with a single pure grandparent. The only piece of 
evidence we have bearing upon this point is the fact 
that we know the hybrid form itself to be capable of 
surviving. 
De Vries has recently published a very interesting 
observation, which ought to be mentioned in this 
connection. It appears that two forms of Cnothera, 
which originated in a cross, and are distinguished as 
