CONCLUSIONS 179 



In the table we find 10,000 fathers (or mothers) arranged 

 vertically in the left-hand column into classes (from V 

 down to v), while the arrays of sons (or daughters) belonging 

 to each class of fathers are given in the horizontal lines. 

 We can thus at a glance read off, not only the numerical 

 proportions of each class of sons belonging to a given clas? 

 of fathers, but also, vice versa, the numerical contribution 

 of each class of fathers to a given class of sons. 



Thus, we see that fathers of the V class have sons : 



Fathers V ( ! ons " R S T U V 

 V. Frequency .. i o 12 10 6 



while, on the other hand, sons V belong to the following 

 fathers : 



Sons V. Per Cent, of Sons. 



35 Fathers V .. 6 .. 17-0 



180 ,, U .. 10 .. 5-5 



671 ,, T .. 10 .. 1-6 



1614 S .. 5 .. 0-3 



2500 ,, R .. 3 .. o'i2 



We see that the sons of the highest talent V are derived 

 from fathers of the classes V to R, all being on the positive 

 side of mediocrity ; and, furthermore, that the percentage 

 of such sons rapidly falls with the decrease of talents in the 

 fathers fathers of the V class yielding 17 per cent, of sons 

 of class V, those of R only o- 12 per cent. Something similar, 

 of course, applies to ah 1 the other classes of fathers and sons. 

 We have, for instance, the same proportions between 

 fathers and sons on the negative side, the lowest class of 

 fathers, v, giving the greatest percentage of sons of that type. 



This definitely proves in figures how important the 

 quality of parents is for the production of quality in off- 

 spring. Coupled with the knowledge that favourable 

 conditions, however much prolonged, cannot alter the 

 hereditary descent of qualities, the conclusion is forced 

 upon us that the only means of eliminating once for all- 



