94 BIOMETRY 



an amount which we may denote by the letter e. 

 The chance that any particular determination differs 

 from the true value by more than twice the probable 

 error is 4-5 to I against. 



The chance that it differs by more than 32 is 21 : i against. 



This is clearly very valuable information to possess 

 when we are dealing with any kind of statistics. 



We must now pass on to consider what methods are 

 available to the biometrician for dealing with the 

 problems of heredity. His way is to take a large 

 number of pairs of relations, each pair consisting, say, 

 of a father and a son, and to find out how much more 

 like the members of such a pair are to one another on 

 the average than the members of similar pairs of 

 individuals would be, if taken at random and without 

 regard to relationships from among the general 

 population to which these fathers and sons belonged. 



Now we shall see later on that this is not the only 

 way of looking at the phenomenon of heredity, nor is 

 it the way which is most familiar to biologists. But 

 it is important to remember that what the biometrician 

 means by amount of inheritance is a numerical value 

 which expresses the average degree of likeness between 

 a particular pair of relatives for example, fathers and 

 sons. 



In the accompanying 'correlation table' a purely 

 imaginary illustration there are tabulated the 

 statures of 4,503 fathers, and those of one son of 

 each of them. Thus 14 fathers, each 62 inches high, 



