104 SEX AND HEREDITY 



above or below the average have sons whose average 

 heights differ from the average of all the sons in the same 

 direction (above or below) as their fathers, but to a 

 smaller degree. And the exact working out of Pearson's 

 figures show that the degree is about one half. In other 

 words, sons on an average inherit half the peculiarity of 

 their fathers. 



A word in passing may be said about a point which may 

 occur as a difficulty to some readers. At first sight it 

 might appear that the fact that tall fathers have sons 

 on an average less tall than themselves, and that short 

 fathers similarly have sons on an average less short than 

 themselves implies that the population is getting more 

 mediocre and less variable in each generation. This, 

 however, is of course not the case, and the reason is seen 

 at once when it is remembered that we are only dealing 

 with average heights. As we saw above, the sons of 

 fathers 65-5 inches tall ranged from 5 feet to 6 feet 2 inches, 

 and in this way, although on the average sons are less 

 peculiar than their fathers, there are always a number 

 of them even more extreme, and thus the supply of tall 

 and short men in the population is kept up. 



We thus see that we cannot predict the height of 

 individual sons of individual fathers, but we can say in 

 general that given a large number of fathers of a given 

 height the average height of their sons will be so and so. 

 And that average height is obtained thus. If the height 

 of the fathers is so many inches above or below the 

 average, the average height of their sons will be about half 

 that number of inches above or below the average. This 1 



1 Readers versed in statistical mathematics will know that 

 properly speaking the prediction of the heights of sons from that 

 of fathers is not obtained directly from the coefficient of corre- 



