364 SEXUAL selection; man. PartIL 



selves and their children, and it is a simple plan to kill 

 their infants. In South America some tribes, as Azara 

 states, formerly destroyed so many infants of both sexes, 

 that they were on the point of extinction. In the Poly- 

 nesian Islands women have been known to kill from four 

 or five to even ten of their children ; and Ellis could not 

 fiod a single woman who had not killed at least one. 

 Wherever infanticide prevails the struggle for existence 

 will be in so far less severe, and all the members of 

 the tribe will have an almost equally good chance of 

 rearing their few surviving children. In most cases 

 a larger number of female than of male infants are 

 destroyed, for it is obvious that the latter are of most 

 value to the tribe, as they will when grown up aid in de- 

 fending it, and can support tliemselves. But the trouble 

 experienced by the women in rearing children, their 

 consequent loss of beauty, the higher estimation set on 

 them and their happier fate, when few in number, are 

 assigned by the women themselves, and by various ob- 

 observers, as additional motives for infanticide. In 

 Australia, where female infanticide is still common. Sir 

 G. Grey estimated tlie proportion of native women to 

 men as one to three ; but others sav as two to three. 

 In a village on the eastern frontier of India, Colonel 

 Macculloch found not a single female child.^^ 



When, owing to female infanticide, the women of a 

 tribe are few in number, the habit of capturing wives 

 from neighbouring tribes would naturally arise. Sir J. 

 Lubbock, however, as we have seen, attributes the prac- 

 tice in^ chief part, to the former existence of communal 

 marriage, and to the men having consequently captured 



^^ Dr. Gerland (' Ueber das Aussterben der Naturvolker,' 1868) has 

 collected much information on infanticide, see especially s. 27, 51, 54, 

 Azara (' Voyages,' &c. tom, ii. p. 94, 116) enters in detail on the motives. 

 See also M'Lennan (ibid. p. 139) for cases in India. 



