INHERITANCE, VARIATION AND SELECTION. 33 



and more easily impressed, and partly because the later impregna- 

 tions would have to share their influence with the previous ones. 

 While the majority of these cases refer to external and apparent 

 characters, those relating to thoroughbred horses refer to functions, 

 and it is quite evident that the functions of organs are fully as liable 

 to be influenced in this manner as are the organs themselves. It 

 is also quite likely that widows have their offspring by a second 

 marriage influenced in the same way. 



Although the statements relating to the effects of previous 

 impregnations are vouched for by many observers, recent inves- 

 tigations have thrown doubt on the whole series of phenomena. 

 The trouble is no one seems to have systematically investigated the 

 subject, and many of the so-called facts, especially those relating 

 to human beings, are open to suspicion. In my examination of 

 the pedigrees of eminent men I have observed what appeared to be 

 an unusual number of widows who married a second time, and the 

 children of those second marriages appeared as progenitors. This 

 is suggestive but not demonstrative, and I have not carried out an 

 investigation along this line because it does not appear that the 

 results would be proof of anything. 



ATAVISM. 



Characters which are ordinarily transmitted from generation 

 to generation sometimes disappear in the child and reappear in 

 the grandchild, the great-grandchild, or even some more remote 

 descendant. This action is called atavism and may be considered 

 as an exception rather that the general rule. There are, however, 

 a good many cases that illustrate this particular action. 



Darwin mentions 19 the case of a pointer bitch which had seven 



(19) Animals and Plants Under Domestication, Vol. II, p. 46. 



