K. Pearson 
141 
It is proper before I go further, to explain how the results for resemblance 
between brothers and sisters of different ages in head measurements have been 
made. In the first place a growth curve for each sex and for every measurement 
was drawn; this growth curve simply consists in plotting the average size of 
head of a child of given age to that age. Diagram V. represents the growth 
DuGBAM V. Oroivth of Auricular Height in GirVs Head. 
(30 
129 
128- 
127 
_c 
126- 
Q 
< 
125 
UJ 
I 
124 
CO 
123 
tn 
o 
122 
o 
121 
h- 
I 
a 
120 
u 
X 
119 
a: 
118 
< 
—1 
117 
O 
116 
< 
115 
H4 
1 
i 
!>-<< 
1 
r 
< 
' i 
5^ 
1 ^ 
r 
) 
i 
A 
i 
V 
A 
\ 
7 8 9 10 II 12 13 14 
AGE of GIRL In YEARS 
15 16 17 
19 20 
of auricular height of head of the mean girl from 4 to 19 years of age. The 
observation points are then smoothed and we obtain the mean growth curve. I 
cannot stay to discuss these mean growth curves now, but it must be clear that 
they give us a method of ascertaining the mean head growth of a child from any 
one year of its life to any other. Now all children do not grow in the same 
manner, but as we are dealing with average results we shall obtain a reasonable 
measure of growth by using the growth curve of the mean child. By means of six 
growth curves like that shown, the length, breadth, and height of every child's 
head was reduced to the dimensions it would most probably have at the age of 
12 years. Thus we were able to compare the likeness in head measurements of 
brothers at the same standard age. This is the method by which the inheritance 
of head length, breadth, and height, given in Table II., was deduced*. 
* By a much more elaborate investigation in multiple correlation I found for resemblance between 
brothers in head length -54 (see R. S. Proc. Vol. 71, p. 294). The growth correlation not being absolutely 
linear, I am not sure that that value is better than the -5 of the present simpler method. 
