INTERNAL CAUSES OF VARIATION 193 



There are certain intermediate cases difficult to name. An 

 occasional cow gives no more milk than her wild progenitor ; a 

 hog resembles not his immediate kind, but his striped and long- 

 nosed ancestor, the wild boar of the bush ; the horse or the 

 sheep paws snow the first time he sees it, though cattle do not ; 

 the dog turns many times around in lying down, as if making his 

 nest ; the occasional horse has bars on his legs and stripes on 

 his shoulders. Which term shall be applied ? 



These are undoubtedly line cases, and all would not agree as 

 to whether they should be regarded as reversions or instances of 

 atavism. Because of our very frequent need for a term to cover 

 experiences met almost every day by the breeder, and due to 

 more near-by causes, the writer is of the opinion that it is better 

 for our purposes to extend the meaning of the word " atavism " 

 well forward, making it cover cases of remote characters, even 

 within the species (like those just given), leaving the term 

 " reversion," which will be much more frequently needed, to cover 

 the more near-by cases that occur every day in our herds, and 

 that would be traceable, could we know all the facts, to an old 

 but not remote ancestor. 



Inheritance is from the race. It is evident that inheritance is 

 not limited to the visible characters of the immediate parent. 

 We constantly forget that every individual is possessed of and 

 capable of transmitting all the characters of the race to which 

 he 'belongs. We forget that his visible characters are not his 

 total possession, but only those which are relatively most prom- 

 inent in his case. Other combinations are easily possible out of 

 the same elements in slightly different proportions, and it is not 

 so strange as we think that a character once in possession of a 

 race tends to persist indefinitely and, perforce, occasionally to 

 become visibly apparent. It is as bound to appear, under the 

 law of chance, as is the one black ball in the box of a thousand 

 or a million, if only throws enough are made. 



The work of Galton, while mostly confined to man, yet shows 

 clearly that inheritance is not in strict line with the visible 

 parental characters, but is in large measure independent of the 

 immediate parents. 1 



1 See the Regression Table, sect, iii, chap. xiv. 



