LO 



ETHEL M. ELDERTON 



The fluctuation is no doubt considerable in our results. But we think that it lies 

 far more in the difficulty of estimating psychical characters, than in any real variation 

 in the degree of resemblance. The fluctuation is greatest precisely in those characters 

 where personal bias and sex bias make the judgment more difficult. 



We are now in a position to compare the intensity of resemblance between cousins 

 with that between brethren. Diagram I, p. 9, shows graphically the distribution of 

 the degree of resemblance of the 70 cases of cousins in this first series and of 

 65 cases of brethren, physical and psychical. The cases were numerically distributed 

 as follows : 



Table II. 



Mean Cousinship : '27 

 Mean Sibship : '51 



■008. Standard Deviation : -093 ± -005. 

 •006. Standard Deviation : -068 + -004. 



An examination of the graph shows that the cousinship group clusters at 25 and 

 the sibship group at '5. These may, we think, safely be taken as working values for 

 cousinship and sibship resemblance for either sex, and we may safely assert that 

 brethren are on the average twice as closely related as cousins. This halving of the 

 degree of resemblance corresponding to the fact that normal first cousins have two 

 common grandparents, whole sibs have four. 



It must be noted : (a) that our data for cousins are not drawn from the same 

 records as those for brothers and sisters. While the Family Records here used for 

 cousins enable us also to deal with brothers, these have not yet been tabled and 

 reduced, except in the one instance of Intelligence. Here the adult brothers gave "54 

 as against the average of three cases of adult male cousins giving '34. Schuster 

 found -56 from adult Oxford graduates for ability. Pearson found '52 for brothers 

 at school and Schuster for schoolboys "56. The half of these fraternal values would 

 be -27, which agrees well with the general cousin average, but not so well with the "34 

 which is a definitely higher value. 



(b) that our data for cousins and sibs are neither from the same records, nor 

 for the same range of characters. In the case of the sibships 21 values were for 

 definitely measurable characters, a much more reliable class of material ; while the 

 consulships of the first series do not present a single measurable character, and only 

 one definitely physical estimate, that of Health. The characters which are common to 



