456 THE GRAMMAR OF SCIENCE 



sons is 70 ".8, or we have a regression towards the mean 

 of the general population. On the other hand, fathers 

 with a mean height of 66" give a group of sons of mean 

 height 68".3, or they hdive progressed towards the mean of 

 the general population of sons. The father with a great 

 excess of the character contributes sons with an excess, 

 but a less excess of it ; the father with a great defect 

 of the character contributes sons with a defect, but less 

 defect of it. The general result is a sensible stability of 

 type and variation from generation to generation.^ 



Now the explanation of this phenomenon of regression 

 is quite simple. Take an exceptional son, for example, 

 of 72" stature, we see that the modal father for sons of 

 this type is 69^9, or on the average an exceptional son is 

 likely to have a less endowed father. The fact is that 

 exceptional parents are very rare, mediocre parents very 

 frequent, and, accordingly, an exceptional individual is 

 more likely to be an extreme variety from less exceptional 

 parents than the product of exceptional parents, simply 

 because of the relatively greater frequency of the former. 

 Now a man is not only the product of his father, but of 

 all his past ancestry, and unless very careful selection has 

 taken place, the mean of that ancestry is probably not 

 far from that of the general population. In the tenth 

 generation a man has 1024 tenth great -grandparents. 

 He is eventually the product of a population of this size, 

 and their mean can hardly differ from that of the general 

 population. It is the heavy weight of this mediocre 

 ancestry which causes the son of an exceptional father 

 to regress towards the general population mean ; it is the 

 balance of this sturdy commonplaceness which enables 

 the son of a degenerate father to escape the whole burden 

 of the parental ill. Among mankind we trust largely for 

 our exceptional men to extreme variations occurring 

 among the commonplace, but, as we shall see later, if 

 we could remove the drag of the mediocre element in 



1 Such differences as we note, e.g. o". 14 ± .11 in type, and o".39 + .08 

 in variability, may well be due to periodic selection acting on sons before they 

 become fathers. 



