282 PROF. K. PEARSON ON THE MATHEMATICAL THEORY OF EVOLUTION. 
Table VIII.—Regression. 
Class. 
Coefficient of Regression. 
Younger sister on elder sister .... 
•4996 
Elder sister on younger sister .... 
•3939 
Younger brother on elder brother . 
•3966 
Elder brother on younger brother . 
•3860 
Sister on brother. 
•3100 
Brother on brother. 
■4547 
Table IX.—Variation in Selected Groups. 
Selected fraternity. 
S.D. 
in inches. 
S.D. 
in inches. 
Unselected. 
Younger sisters of selected elder . 2'2512 
Elder sisters of selected younger . P9989 
2-5119 
2-2303 
All younger sisters of sisters 
All elder sisters of sisters 
Younger brothers of selected elder . 
Elder brothers of selected younger . 
2-4327 
2-4000 
2 6434 
2-6080 
All younger brothers of brothers 
All elder brothers of brothers 
Sisters of selected brother 
Brothers of selected sister .... 
2-0789 
2-5178 
2-2430 
2-7164 
All sisters of brothers 
All brothers of sisters 
These tables have been calculated in precisely the same manner as the previous 
series. 
(b.) Conclusions .—Now these results seem at several points of very great sug¬ 
gestiveness. In the first place, with regard to variation, we see that elder sisters are 
significantly more mediocre than younger sisters ; younger sisters are taller and more 
variable. The same difference appears in the case of elder and younger brothers, but 
the probable errors do not allow us in this case to assert that the difference is 
certainly significant. To illustrate this conclusion we give the constants for pairs of 
sisters, no respect being paid to relative age. 
