On a Measure of the Resemblance of First Cousins in Man. 



J 



By Ethel M. Eldertox, Galton Research Scholar in National Eugenics 

 in the University of London ; assisted by Karl Pearson, F.R.S. 



1. Introductory. While a very large amount of data has been collected, 

 reduced and published relating to the degree of resemblance in physical and psychical 

 characters of a considerable number of pairs of relatives in man — especially in the 

 direct line and between collaterals of the first degree — but little has yet been done 

 with regard to collaterals in higher degrees. As far as we are aware tbe only quanti- 

 tative measures yet determined are those for eye colour in man between uncle or 

 aunt and nephew or niece*. No measure of resemblance has yet been determined for 

 cousins. Yet it is precisely among collaterals of the second degree that the question 

 of consanguineous marriages becomes in practical life of great importance, inequality 

 of age being bere less marked, and thus the degree of resemblance between such 

 collaterals has not only scientific but eugenic value. According to local law and 

 religious custom cousin marriages are permitted or forbidden ; thus it would appear 

 that we are here concerned with divergent human experiences, unconsciously formu- 

 lated, as to the relative value of endogamy and exogamy. If we take a character 

 whicb is detrimental to tbe individual, it will, at least in primitive communities, 

 be in the bulk of cases a hindrance to mating. Hence, as a rule, we must classify 

 such a detrimental character as recessive in the Mendelian senset, otherwise selection 

 would have weeded it out. Now consider for a moment a population of dominants 

 with notation DD, and suppose one of these to mate with an individual of detrimental 

 attribute and constitution RR. The result will be the hybrid sibship marked by DR, 

 in which the recessive character R will be latent. If brother-sister mating is for- 

 bidden the next generation will be obtained (assuming the recessive individuals RR 

 to be extremely rare) by mating with the population of dominant character, and the 

 result will be equal numbers of DD and DR. Thus the generation of cousins would 

 consist of 50 p. c. of dominants and 50 p. c. apparent dominants with the detrimental 

 character recessive. ' It therefore follows that it would be as detrimental for some 

 cousins to marry as for all brothers and sisters of the first hybrid sibship. That is to 



* Pearson and Lee : Phil. Trans. Vol. 195, A, p. 114 et seq. 



t Thus albinism and the tuberculous diathesis are, if not true and complete recessives in the Mendelian 

 sense, still more nearly recessive than dominant characteristics. But there are other abnormalities, e.g. 

 certain digital deformations, which are nearer to, if perhaps not true, dominants. 



1—2 



