The Decisive Experiment 175 



for Weismann's amputation which can prove nothing. 

 Similarly light hammering, continually repeated, which 

 a proper mechanism might automatically perform upon 

 certain parts of the skull of hornless animals, would be 

 better than cutting off or breaking the horns in horned 

 animals. 



All these artificial stimuli would certainly produce in 

 each individual the hypertrophy of the organ upon which 

 they act. It could then be seen whether the repetition of 

 these stimuli throughout a series of generations would be 

 followed by the production of individuals in which these 

 organs would possess at birth even in a small proportion 

 the greater development that had been acquired in several 

 successive generations of its ancestors. The performance 

 of such experiments upon guinea pigs or rats would not 

 seem to present very great practical difficulties; never- 

 theless, so far as we know, it has never yet occurred to 

 any one to make them or to propose them. 



But in all these experiments one must never forget 

 that it is just the littleness of the inheritable fraction of 

 an acquired quantitative variation, that constitutes the 

 great difficulty of verification of the Lamarckian principle. 



Galton proposes as is known to select for experi- 

 ment rather the inheritance or non-inheritance of certain 

 acquired instincts. He advises for example to adopt the 

 method of the following experiment of Mobius upon the 

 pike; Mobius divided a large glass receptacle into two 

 compartments by means of a perfectly transparent glass 

 septum and placed the pike in one compartment and in 

 the other little gudgeons upon which the pike usually 

 feeds. It followed that whenever the pike precipitated 

 himself toward any of the little fishes he was stopped by 

 the glass against which he hit. After several weeks of 



