122 Relationship of Tntelligence to Size and Shape of Head 
both boys and girls the slow increase of intelligence with size of head is quite 
sensible, but the variability of each group is such that some 25 per cent, of the 
dull have larger heads than the very intelligent, and some 25 per cent, of the very 
intelligent have heads smaller than the dull average. It is again impossible under 
such conditions to use head size as a basis for judgment as to intelligence. Of 
course all this is merely stating in other words what is obvious to the biometrician, 
when he finds a low correlation, but possibly, although I am hardly hopeful, it may 
help to convince the anatomist and old school anthropologist that head measure- 
ments are not of real service as intelligence tests*. 
In Table IX are placed the mean and variability of each array of head 
measurements corresponding to different intelligence groups in the case of the 
Canibridoe graduates. In Table X we have the same data for the school children. 
; 
1 1 
1 1 
1 1 
1 1 
\ \ 
1 1 
i 1 
1 1 
1 1 
„ Mead Length 
1 1 
1 4 
z ' 
< i 
Z ' 
Z 1 
< 
1 
z \ £i 
^' } 
4 <l 
Mean 
. 1 
Group 
a. 
0 
£C 
O 
a ' u' 
D 1 Z| 
o u' 
O , ' 
1 
Group 
1 1 
1 ol 
« 1- 
' 1 
1 
1 h^EAO Breadth 
1 
\ ^"1 
\ h 
! 
1 1 
1 1 
1 1 
1 1 
1 1 
1 1 
1 1 
1 1 
I"' Class Honours 
2"rCL,«ss 
HofNOUHS 
3"" Class 
Honours 
Pass Degrees 
Fig. 5. 
* Nurture, exercise and nourisliment — shortly environment and class — district or local race, influence 
extensively the anthropometric measurements. We cannot compare pauper imbeciles or hospital post- 
mortem results with middle class students or professors. We cannot measure agricultural labourers and 
men of science and point triumphantly to great differences in head volumes as marking widely separate 
intellectual grades. See the Ih-itis), Mnlical Journal, March 3, p. 530, and March 17, p. G51, 1900. 
