CH. IIL] ORIGIN OF TASTE FOE SCIENCE. 145 



experience will have shown them how surely, in 

 every case with which they have dealt, the 

 great majority of causes, or what might be better 

 named " nre-efficients," admitted of being ana- 

 lysed and grouped into natural orders, leaving a 

 minority of unclassed influences, which them- 

 selves form a class of their own, and which can 

 be reduced indefinitely, in proportion to the 

 minuteness with which the statistician cares to 

 pursue his analysis. The statistics of railway 

 accidents will serve as an example. When Cap- 

 tain Douglas Galton was secretary of the railway 

 department of the 'Board of Trade, he succeeded 

 in sorting their causes into the groups in which 

 we have since been accustomed to see them 

 printed year after year. So long as the general 

 system of management of a railway is little 

 changed, the same statistical ratio is maintained 

 among them, a given proportion of accidents being 

 due to this cause, and another to that. We may 

 therefore estimate with some certainty the saving 

 of life and limb, or of material of various descrip- 

 tions, that will be effected when any one of these 

 causes shall be wholly or in part removed. Shni- 



L 



