568 Species According to the Theory of Mutation. 
a much more complicated structure and have arisen in 
a historical way. They cannot be isolated and then sub- 
jected to experiment like chemical bodies; we can only 
investigate them by studying the relation of closely allied 
species and varieties, i. e., forms in which a definite unit, 
or several of them, are present in one plant and absent in 
the other. For this reason our investigations are, for 
the present at any rate, confined to the units which have 
arisen most recently. 
But even as it is the business of comparative science 
in general, first, to apply the conclusions derived directly 
from the facts, to cases that have not been themselves 
observed and then to extend them gradually further and 
further, it is our duty and our right to test the applica- 
bility of our conclusions as thoroughly and as widely as 
possible. 
Therefore we have now to face the question whether 
the theory of the origin of species by mutation and the 
theory that hereditary characters are composed of ele- 
mentary units are in harmony with the theoretical concep- 
tions to which systematic science on the one hand and 
embryology on the other have given rise. If it can be 
shown that the mutation theory satisfies the demands of 
these sciences better than the present form of the theory 
of selection, its justification as a theory of the nature 
of inheritance will, in my opinion, be placed on a sure 
foundation. 
For this reason I shall devote the last part of this 
work to general considerations of this kind. In doing 
so I leave the safe ground of facts and venture into a 
region in which I can no longer mainly depend on my 
own experience. But experimental inquiry must derive 
its problems from this more general aspect of the ques- 
