36o Darwinism and the problems of life 



man of science has denied this. The transformation of 

 one species into another can only be established with 

 certainty in a few cases — such as that of the Porto- 

 Santo rabbit. That is natural enough, as this trans- 

 formation lasts far too long and is far too gradual to 

 come under human observation. We may, therefore, 

 readily grant that we cannot establish the selective value 

 of variations in particular cases. We cannot do this 

 solely because we do not know what is of value in the 

 life of the animal. And if that is impossible with living 

 animals the conditions of whose life are known to us, how 

 can we be expected to show the selective value of 

 variations in animals that have evolved to their present 

 forms in earlier ages. We do not know the accidents — 

 the isolations — that determined selection to modify them 

 in a particular direction, and so cannot know what has 

 selective value for this particular direction. All that we 

 can say is that a variation will have selective value if it 

 favourably influences the strength of the particular 

 animal, as that will lead to an increased multiplication 

 which must be gradually prepared. It is clear, at all 

 events, that there must be a large number of small 

 variations with selective value, and that it is incorrect 

 to say that there are hardly any variations with selective 

 value. 



The action of natural selection cannot be directly 

 observed in nature, we must admit. But this admission 

 does not cover a defect of the theory of selection. 

 There are many theories the truth of which cannot 

 be established by direct observation. It is said that 

 light is due to the vibration of the|^ minute particles 



