No. 627] 



ADAPTATION 



which variations are produced and not in the fact that 

 many of these variations fail to maintain themselves. 



This argument is so plausible that it seems self- 

 evident. And indeed in a sense it is. But there is an- 

 other sense in which it is quite specious. Truly enough, 

 no individual can survive which is not first born or 

 hatclied, or in some way brought into being by its par- 

 ents. And those peculiarities which distinguish one in- 

 dividual from another are largely ushered into life along 

 with it. They exist prior to selection. But fitness is a 

 relation, not an absolute proi)erty of the organism. The 

 word denotes merely a ('(■l iaiii iiicasnre of adjustment to 

 specific conditions of lite and tlir degree of this adjust- 

 ment we know to vary almost iiuldinitely. To say that 

 the conditions of life, acting through the selective proc- 

 ess, can not be the cause of an increasing degree of fitness 

 is like denying that a sculptor produces a statue, on the 

 ground that he does not create the stone. It is well to 

 note that even the sculptor's function is wholly selective. 

 He eliminates certain portions of an unsliaped mass of 

 material.^^'* 



The foregoing analogy admittedly fails in one impor- 

 tant respect. It implies that the possibilities of selec- 

 tion in a given race are wholly unlimited. We know this 

 to be very wide of the truth. The question to be an- 

 swered here is merely whether or not they are completely 

 random in the sense which has been employed through- 

 out this article. 



Now, some selectionists are wont to deny the com- 

 pletely random character of variation. So far as this is 

 simply a denial of the infinite variability of any species, 

 it is a mere truism. We may perhaps admit the possi- 

 bility that a given strain might, through rigid selection, 

 acquire the "habit" of varying preponderantly in cer- 

 tain definite directions, thus limiting the possibilities of 

 further evolution witliin that group. And we might even 

 grant that sucii defuiiicly directed variatii^iis might ac- 



