SEXUAL SELECTION IN RELATION TO MAN 757 



the world, including therein the intermarriage of brothers 

 and sisters. The late Sir A. Smith, who had travelled 

 widely in South Africa, and knew much about the habits 

 of savages there and elsewhere, expressed to me the strong- 

 est opinion that no race exists in which woman is considered 

 as the property of the community. I believe that his judg- 

 ment was largely determined by what is implied by the term 

 marriage. Throughout the following discussion I use the 

 term in the same sense as when natutalists speak of animals 

 as monogamous, meaning thereby that the male is accepted 

 by or chooses a single female, and lives with her either dur- 

 ing the breeding season or for the whole year, keeping pos- 

 session of her by the law of might ; or, as when they speak 

 of a polygamous species, meaning that the male lives with 

 several females. This kind of marriage is all that concerns 

 us here, as it suffices for the work of sexual selection. But 

 I know that some of the writers above referred to imply by 

 the term marriage a recognized right protected by the tribe. 

 The indirect evidence in favor of the belief of the former 

 prevalence of communal marriages is strong, and rests chiefly 

 on the terms of relationship which are employed between 

 the members of the same tribe, implying a connection with 

 the tribe, and not with either parent. But the subject is 

 too large and complex for even an abstract to be here given, 

 and I will confine myself to a few remarks. It is evident 

 in the case of such marriages, or where the marriage tie is 

 very loose, that the relationship of the child to its father 

 cannot be known. But it seems almost incredible that the 

 relationship of the child to its mother should ever be com- 

 pletely ignored, especially as the women in most savage 

 tribes nurse their infants for a long time. Accordingly, 

 in many cases the lines of descent are traced through the 

 mother alone, to the exclusion of the father. But in other 

 cases the terms employed express a connection with the tribe 

 alone, to the exclusion even of the mother. It seems possi- 

 ble that the connection between the related members of the 

 same barbarous tribe, exposed to all sorts of danger, might 



