THE PARALLEL GROWTH OF OPPOSITES 29 



an artificial environment is that nature demands from the 

 creature a nice adaptation in all its parts, which cannot 

 be disturbed without the gravest danger to its continued 

 existence. This principle applies when the environment is 

 regarded as stationary, as Mr. H. Spencer sometimes appears 

 to conceive it. In hypothetical conditions of that kind 

 variation does not seem to be called for at all, and excessive 

 variation would certainly be fatal. It is equally valid in 

 a changing environment, such as that we are actually 

 acquainted with. Excessive variation in that produces 

 monsters, which, like the products of artificial breeding, 

 are violent deviations from an established type, when there 

 has been no corresponding variation in the environment, 

 and for which, in the same way, the prospects of survival 

 are slight, or none at all. 



This delicate equilibrium between the different parts 

 of an organism, and between the whole organism and its 

 environment, is at least as necessary to survival in the 

 highest forms of life as in the lowest. Indeed, it is likely to 

 be more so, inasmuch as increased complexity is attended by 

 increased liability to derangement. A watch is more likely 

 to go out of order than a ploughshare. If this be true, 

 increased complexity, instead of being the criterion of fitness 

 to survive, is rather the reverse, and tells against the chance 

 of survival ; and the a priori probability appears to be, to 

 some extent, confirmed by the comparative permanence 

 of more and less highly developed types in the past, which 

 has already been referred to. 



Simplicity of structure is not, then, the criterion by which 

 the destructive forces of nature are guided in their selection, 

 and, as far as I know, no other has ever been suggested. 

 No single characteristic of form or function, or assemblage 

 of such characteristics, can be indicated as having served 

 in the past, or being likely to serve in the future, as a con 

 stant source of danger to the creature possessing it. To 

 say that the unfit are eliminated is mere tautology, unless 

 some general quality can be detected which has always served 

 as a criterion of unfitness. If we were able to distinguish 



