Essays in Statistics. 



189 



the judges of the men whose kinships the author has studied 

 most closely with the last column, that which gives the averages, 

 that is, with the column which purports to express the law in 

 numerical terms. 



The number of families that has served as the basis of the work 

 is about 300, and includes nearly 1000 men of note, of whom 

 415 are illustrious. The author thinks that, if there is a law, so 

 great a mass of facts ought to bring it to light. This law is given 

 in the last column of Table II. The probability that a man of 

 mark would have remarkable kinsmen is, for his father, thirty-one 

 per cent. ; brothers, forty-one per cent. ; sons, forty-eight per cent, 

 etc. (See Table II., column 9.) 



If we estimate the probability of the kinsmen of illustrious men 

 rising to be eminent and the author shows that eminent men are in 

 general less numerous by one half than illustrious men it will be 

 found to be as follows : 



In the first degree, for the father as one to six ; for each brother 

 as one to seven ; for each son as one to four. In the second 

 degree, for each of the grandfathers, as one to twenty-five ; uncle, 

 one to forty ; nephew, one to forty ; grandson, one to twenty-nine ; 

 In the third degree, for each cousin-german, one to one hundred ; 

 each of the other relatives one to two hundred. 



Before we dismiss statistics we must clear up one point. In 

 Table II. the word * father ' stands for ' mother,' as well, and 

 'brother* includes 'sister'; in a word, the male and female rela- 

 tives are indicated by one term. We have now to determine the 

 respective positions of the males and females in the eight groups 

 of one hundred families each. 



TABLE III. 



