THE PARALLEL GROWTH OF OPPOSITES 33 



ball, or a sheep, or a sugar-cane, he knows with fair exact 

 ness what he wants. The second kind of improvement 

 has no terminus or ideal, but is merely the continuation 

 of an endless process. The processes of artificial selection 

 differ from the processes of nature in this vital character 

 istic that with the former we know what end is aimed at, 

 with the second we do not. 



Remembering, then, that every organism depends for its 

 continued existence on a nice adjustment between its 

 constituent parts and its environment, and that this pro 

 position is equally true of every organism in all stages of 

 evolution, we may go on to consider in what sense, if any, 

 nature can be said to make a selection when one class goes 

 under, and another remains, and still beholds the light of the 

 sun. In order to do this, we must have a clear idea of what 

 we mean by selection. 



Selection, like improvement, is a concept borrowed 

 from the furniture of the human mind. When applied to 

 nature its use is anthropomorphic, and it cannot mean 

 anything essentially different from what it means when 

 applied to man. Even if the use is only by analogy, there 

 still must be some points in common. Now the minimum of 

 connotation for the term is the differential treatment of two 

 objects. In the second place, the two objects between which 

 the choice is made must themselves differ or have different 

 implications ; if two objects are exactly alike, they must 

 be equally desirable or undesirable, and choice is impossible. 

 In the third place, in order that choice may have any scientific 

 value or meaning, it must be based on some known criterion. 

 If a blind man selects one of two pictures, his choice is for 

 tuitous ; the next time he would probably select the other. 

 Again, if we want to make any scientific use of our obser 

 vations, we must know what the criterion is. Unless the 

 criterion is known, the preferential action is no better for 

 our purposes than a blind choice. Finally, in men at least, 

 there must be free-will. This, of course, is not found in 

 nature; but, if the other constituents in the concept of 

 human selection were found in the processes of nature, it 



BBNETT C 



