224 Experimental Zoology 
Evolution by Means of Definite Variation 
The evidence which we now possess indicates, with some 
probability, I think, that in some cases, at least, the process of 
evolution may have been by definite or discontinuous changes 
from one fixed form to another fixed form. The essence of the 
process is not that the change has been marked or great, but 
rather that the new type is, from its first appearance, a definite 
step in a new direction. What has been gained does not need 
to be maintained by any process of selection, but becomes a 
part of the permanent inheritance. It is probable that it may 
be found convenient to distinguish between different kinds of 
mutations. In fact, de Vries has already distinguished between 
mutations that involve only a loss of a character (leading to the 
formation of a “‘variety’’ —a varietal or retrogressive mutation), 
and mutations that are new and form new elementary species — 
progressive mutations; and mutations that are due to a latent 
character reappearing —degressive mutations. Of greater im- 
portance for the theory of evolution would be the distinction 
between those progressive mutations that become recessive in the 
first generation, when crossed with the parent forms, and subse- 
quently split according to the Mendelian proportion, recessive 
mutations ; those mutations that dominate in the first generation 
after a back-cross, and may or may not split subsequently, 
dominant mutations; and those that produce a new permanent 
form or fixed hybrid, when crossed with the parent hybrid 
mutations. Any one of these three kinds of mutations may give 
the starting point for a new group of organisms, a new species, 
or a new variety, but the chance that they may do so will be dif- 
ferent in different cases, and this important consideration must 
next be examined. . 
It is recognized that one of the greatest difficulties that Dar- 
win met with in his theory of natural selection was the swamp- 
ing effects of crossing. If a new and better fluctuation were 
to appear, it would, in most cases, in order to perpetuate itself, 
have to unite with the parent form. It will be recalled that 
