ILLUSTRATION OF STATISTICAL RESULTS 331 



a gradual sinking back towards the mean height of the popu- 

 lation. This general conclusion is in accordance with the 

 experience of breeders (in non-Mendelian cases), who, having 

 temporarily secured a high-class breed by persistent artificial 

 selection, find that there is a gradual deterioration during a 

 period of inbreeding without selection. 



The importance of definite conclusions of this kind can hardly 

 be overestimated. 



" Looked at from the social standpoint, we see how exceptional 

 families, by careful marriages, can within even a few generations 

 obtain an exceptional stock, and how directly this suggests 

 assortative mating as a moral duty for the highly endowed. 

 On the other hand, the exceptionally degenerate isolated in the 

 slums of our modern cities can easily produce permanent stock 

 also : a stock which no change of environment will permanently 

 elevate, and which nothing but mixture with better blood will 

 improve. But this is an improvement of the bad by a social 

 waste of the better. We do not want to eliminate bad stock 

 bj' watering it with good, but by placing it under conditions 

 where it is relatively or absolutely infertile " (Pearson, G^ammay 

 0/ Science, p. 486). 



By statistical methods Pearson has reached the interesting 

 conclusion that while blended inheritance illustrates regression, 

 it is to cases of exclusive inheritance that we should look for 

 reversion {i.e. the reappearance of a character which occurred 

 in a definite ancestor). In exclusive inheritance, in which the 

 offspring inherits the full character of either parent, and does 

 not blend the two, the law of ancestral inheritance in the strict 

 sense ceases to hold, for it presupposes a blend. Thus eye- 

 colour in man rarely, if ever, blends, and it is in regard to such 

 characters that we should look for reversion. 



As a matter of fact, Pearson notes that of exclusive inheritance 

 with reversion he has as yet (1900) discovered no case, except 

 possibly coat-colour in dogs. He gives a list of various pheno- 



