200 HEREDITY AND EUGENICS 



number of births was 13,360,557. The frequency of 

 twin births was i in 89-1, and of triplets i in (88-9)^. 

 For quadruplets, however, there is an excess over 

 expectation, the number being i in (71 -9)^. Again, 

 in the United States birth-registration area in 191 7 

 the number of births is recorded as 1,339,975, the 

 number of twins being i in 93-1, and of triplets i in 

 (93-0)2. From these statistical relations it would 

 follow that " triplets are produced by the coincidence 

 of two independent processes occurring with equal 

 frequencies." One of these processes gives rise to 

 twins. This relation w^ould hold whether multiple 

 births occurred through multiple ovulation or budding 

 of a monoz3^gotic embryo or both, provided that each 

 followed the rule. This result is remarkable when 

 one considers the number of conditions, both bio- 

 logical and human, which contribute to produce the 

 registration statistics. 



Man appears to be unique among animals, as 

 Davenport points out, in that he produces both types 

 of multiple births: (i) by the budding method found 

 in the armadillo; (2) multiple births from separate 

 eggs, as found in carnivora, herbivora, and rodentia 

 generally. Differences in the degree of resemblance 

 betw^een identical twins may depend in part on the 

 stage at which budding of the original embryo takes 

 place. Much more information is required regard- 

 ing human multiple births, particular^ data from 

 obstetricians on such matters as the number of 

 chorions present, the presence of blighted embryos 

 in the afterbirth, and the relation between the number 

 of corpora lutea and the number of embr3^os at a 

 birth. The study of twins presents many sides, 

 and will furnish evidence on a variety of questions 

 connected with heredit}^ and reproduction in the 

 human race. 



The recent study of the so-called free-martin or 



