122 THE THEORY OF MUTATION 
result of Pearson’s calculations, described in the 
preceding chapter, seems to indicate that the selection 
of a certain value of a particular character for many 
generations will never lead to the formation of a race 
in which the mean value of the character is as high 
as the selected value. But, says the selectionist, it 
will happen in Nature that as the standard of the 
race is raised by selection, the value selected will be 
still further raised, and so on, and in this way an 
indefinite amount of improvement is rendered possible. 
If Johannsen’s conclusions are well founded, this is 
clearly not the case; on the contrary, there is a 
perfectly definite limit to the effect which selection 
can produce. 
The question whether or not a gradual method of 
evolution is possible has not yet been absolutely 
decided for any single species or character, but it cer- 
tainly seems that now for the first time the possibility 
of a definite decision is within sight. At the same 
time it is impossible to prove a universal negative. If 
we look at the other side of the problem we shall find 
that the evidence in favour of an alternative process 
has multiplied even faster than the evidence against 
the continuous accumulation of minute differences ; 
and the present tendency is certainly to look for other 
sources of specific distinctness than that which is 
offered by the natural selection of continuous varia- 
tions. 
Even before the new evidence which we have briefly 
outlined was available, Herbert Spencer found the 
difficulties in the way of accepting the purely 
