504 



ON THE ESTIMATION OF 



TABLE III. 



(9.) Although the general agreement of the results in 

 each horizontal line of this table is quite sufficient to 

 afford a highly satisfactory verification of the theory, 

 it is impossible not to be struck with the uniform 

 and steady increase (though small) in the value of a 

 in proceeding from the gold outwards. There occurs 

 not throughout the whole table a single instance in 

 which this progressive dilatation is not maintained. For 

 this there must be a reason ; and the only rational 

 account of it, so far as I can perceive, is this : Were the 

 number of shooters infinite, including (indifferently) every 

 gradation of skill, from absolute random shooting up to 

 absolute certainty of striking the point aimed at : the 

 result of their combination would be one from which in- 

 dividual skill would be entirely eliminated ; and the 

 distribution of the hits would be regulated purely and 

 simply by the intention of hitf ; ng the centre. The skill 

 in that case (on which the value of a depends) would be 

 the average skill of the whole human race ; and the value 



