. 
212 The American Naturalist. [March, 
cusp in one jaw developing, while the cusp in the other jaw, 
opposing it and presumably stimulating its development, is degen- 
erating.” The force of these exceptions will weigh seriously 
against the Lamarckian principle, unless they also can be proved 
by subsequent research to conform to the laws of individual 
adaptation. I consider that the strongest line of attack which 
can in futute be taken against Lamarckism will be in showing that 
certain characters (such as the above), in which it is supposed to 
operate, could not be produced on principles of direct adaptation. 
But if we reject the Lamarckian principles we must assign 
Selection as the cause of these definite lines in variation, for no 
one would urge the third alternative. 
6. What is the Relation between Variation and Selection ?— 
The question of Utility is the first which arises when we attempt 
to explain the origin of such variations as we are here consider- 
ing by the selection principle. In the recent animated discussion 
which has taken place between Romanes," Mivart, and others on 
the one side, and Wallace® and Dyer on the other, great difference 
of opinion has been shown as to Utility, So far as the question 
bears upon the substitution of pure natural selection for Lamarck’s 
principle, we may, in this argument, avoid the broader question 
by admitting that all characters possess, or have once possessed, 
some degree of utility, or the reverse. This is as necessary for 
Lamarck’s as for Weismann’s principle. The essential question 
here is whether the plus- or minus-variations in advanced stages, 
or the variations in initial stages, or still more the variations which 
constitute the initial stages themselves, are of such importance as 
to weigh sufficiently in the scale of survival, to accumulate defi- 
nite lines of adaptive variations.” Let us assume that they can 
be, what further assumptions are necessary ? 
We start with the proposition that all these variations have their 
origin under the laws which we have seen govern variations of 
Classes I. and II., for upon Weismann’s principle we cannot admit 
51 I refer to the paraconid and cage 
58" The majority of specific inutile (non-adaptive).”” Nature, '89, P- 8. 
r 
Ta y PORRE £ eh 
59‘ There is no p t ific ch fi 
1 j aatas: e 
6 Darwin distinctly dindoun the utility priit t in ite case of Saturnia. See letter 
to Moritz Wagner, “ Life and Letters,” Vol. HI. 
* 
