DEVELOPMENT BY NATURAL SELECTION. Ill 



referred back for explanation from delicacy of structure 

 to constancy of environment. But evidently the con- 

 stancy in the environment in immediate contact with 

 an organism depends on the organism itself. So we 

 are again referred back to organism, and the upshot 

 of the attempted mechanistic explanation is vain. 

 Structure, environment, and activity are associated in 

 what we recognise as a manner normal to the organism ; 

 but of this association we can conceive no possibility 

 of a mechanistic account ; and we are no better off 

 if we return to the vitalistic theory still held by a 

 minority of biologists. From the mechanistic stand- 

 point we are thus face to face with a mystery which 

 is just as insoluble in connection with all physiological 

 activity as the mystery of heredity. In both cases 

 we are up against hard facts which cannot be reconciled 

 with any mechanistic conception. 



It is impossible to ignore these facts. They come 

 before us at every point in biological investigation. 

 They signify perfectly clearly that in the phenomena 

 of life we have before us a kind of unity which cannot 

 be resolved into simpler elements. This unity of 

 structure, environment, and activity is just life. We 

 cannot resolve life into mechanism, because when wc 

 separate a living part from its environment, or suspend 

 its activity, it is at once completely altered. We 

 can take a machine to pieces, or stop its action, with- 

 out evident alteration to the parts ; but we cannot 

 do this with a living organism. Moreover, the unity of 

 a living organism persists : if this were not so, the 

 normals of the organism would not be normals. We 

 must thus recognise that when we say that biology 



