CHAPTER XXXVI 



THE STATISTICAL STUDY OF HEREDITY 



IN MAN 



Introduction. — A good example of this method is afforded by 

 Popenoe and Johnson's studies of the heredity of longevity (chap, 

 xxxix). Other examples are found in Heron's and in Orr and Rosenoff's 

 studies of the heredity of insanity presented in the last chapter. 

 The proof of a given character being hereditary is believed to be at- 

 tained when it is shown that affected parents have a significantly 

 larger proportion of affected offspring than have non-affected parents. 



The statement that twinning is hereditary is based on the fact that 

 twins are far more common in certain strains than in the general popu- 

 lation. A twin birth occurs in about i out of every 88 confinements. 

 C. H. Danforth found that in the case of 50 newborn twins there were 

 171 single brothers and sisters and 10 pairs of twin brothers and sisters, 

 a ratio of about 1 to 18, as compared with that in the general popula- 

 tion, 1 to 88. Moreover, with respect to the mothers of these twins, 

 their brothers and sisters numbered 318 singles to 10 pairs of twins, 

 a ratio of 1 to 32. With respect to the fathers of these twins, there 

 were 219 brothers and sisters born singly and 8 pairs of twin brothers 

 and sisters, a ratio of 1 to 37. A number of similar studies give com- 

 parable results, showing that twinning has a hereditary basis, but in- 

 dicating nothing as to the exact mode of heredity. 



The strength of heredity between different groups of relatives 

 has been determined for a large number of human characters by de- 

 termining the coefficient of correlation (see chap, xliii) existing be- 

 tween selected groups. Complete correlation is represented by the 

 unit, 1. If correlation is 1, heredity is perfect, and various grades of 

 imperfect heredity are represented by decimals, such as 0.9, 0.6, 0.25, 

 etc. It has been shown that for most characters the correlation of 

 offspring to fathers is about 0.4, the same to mothers. The correla- 

 tion between sibs (brothers and sisters) is about 0.5, that between cous- 

 ins about 0.25, that between one-egg (identical) twins 0.9+. These 

 indices of correlation run strictly parallel with degrees of closeness of 

 genetic relationship. 



470 



