30 ENGLISH MEN OF SCIENCE. [CHAP. 



separate qualities are treated in the table. As 

 regards height, there is a stricter method of in- 

 vestigation, which statisticians will appreciate. 

 It is well known, by repeated experience, that 

 the heights of men and of women in any large 

 group are distributed according to the " law of 

 frequency of error." In other words, the propor- 

 tionate number of people of different heights 

 corresponds to what would have been the case 

 supposing stature to be due to the aggregate 

 action of many small and independent variable 

 causes. The probability is inconceivably small 

 that all the independent causes should in any 

 given case co-operate to produce an excess of 

 height; if they did so, the result would be a 

 Brobdignagian giant ; or that they should all 

 co-operate to produce a deficiency in height, in 

 which case the result would be a Lilliputian 

 dwarf. On the other hand, the probability is 

 great that the number and effects of the causes 

 in excess and those in deficiency of their several 

 average values will be pretty equal. As for these 

 and all other intermediate cases, their relative 

 frequency is determined by the above law, which 



