EVOLUTION 489 



Now we have already defined exclusive inheritance as 

 one in which the offspring inherits the full character of 

 either parent and does not blend the two. We do not 

 understand by this that all the offspring will take after 

 one parent ; some may take after one and some after the 

 other. Under such a form of inheritance we shall find 

 it easy to investigate whether reversion is occurring, for 

 occasionally we shall have offspring inheriting from neither 

 parent, but displaying the character of a grandparent after 

 whom neither parent has taken. It will be at once clear 

 that for such inheritance the law of ancestral heredity 

 ceases to hold ; it might express the proportions of re- 

 version, it cannot give the proportions of a blend, for such 

 no longer exists. Let us consider the theory first in its 

 simplest form, namely, that in which there is no reversion, 

 but every child is equally likely to take after its mother 

 or its father. In this case, if we are seeking parental 

 correlation, we should expect to find one-half the offspring 

 identical in character with, say, the male parent, and 

 thus having perfect correlation with him ; the other half 

 would have no correlation with him at all, but would, 

 supposing no assortative mating, be, as far as he is con- 

 cerned, a purely random selection. We haVe thus a 

 mixture of 50 per cent of uncorrelated and 50 per cent 



writers " — a statement we have amply verified in the last section — he means 

 regression. By reversion I denote the full reappearance in an individual of 

 a character which is recorded to have occurred in a definite ancestor of the 

 same race. For example, the father, mother, and three grandparents of a 

 man have brown or black eyes, the man and one grandparent light-blue eyes. 

 By atavisjn I understand a return of an individual to a character not typical 

 of the race at all, but found in allied races supposed to be related to the 

 evolutionary ancestry of the given race. For example, supplementary mammce 

 on the breast of a woman comparable with the two pair on the breasts of the 

 Lemurs, emphasised projection of the canine teeth in man in the same manner 

 as in the anthropomorphous apes, etc. Those cases Darwin speaks of as 

 reversion, although there may be no immediate ancestral history such as there 

 so often is in the case of polydactylism in man. Thus I look upon atavism 

 as an abnormal variation with no immediate ancestral history ; reversion as 

 not necessarily an abnormal variation, but always having an immediate an- 

 cestral history. It is quite possible, of course, for one to pass into the 

 other. But in reversion we are considering a variation, normal or abnormal, 

 from the standpoint of heredity in the individual ; in atavism we are con- 

 sidering an abnormal variation from the standpoint of the ancestry of the 

 race. 



