118 Relationship of Intelligence to Size and Shape of Head 
slightly older ; the lower approximately vertical broken line shows that age is very 
little dependent on the mental class in boys, when we exclude the very intelligent 
group. The approximately horizontal broken lines show the direct influence of age 
on intelligence. Their close approach to horizontality indicates how slight is the 
relationship ; the variations are extremely small as compared with the whole 
mental range. But we see the parallelism of the variations in the two sexes; 
the general changes being shown in an exaggerated diagrammatic manner in Fig. 3. 
The general results reached in this section, for example, the correlation ratios, 
are not dependent on the choice of a normal distribution scale, but that scale 
enables us to plot our results in a manner which indicates conveniently their 
graphical validity. A further graphical illustration is given in Fig. 4. Here 
■VARIATION 
IN 
— A BlU 1 TY- 
5oa 
<joo 
20O 
100 
SCO 
too 
300 
loo 
TOO 
INDI 
20 
ftL S 
-Bo 
(fS — 
r 
< 
J 
u 
! 
cr 
o! 
-s 
QUICK INTELLIOENT 
INTELLIGENT 
SLOW 
INTELLIOENl 
SLOW 
SLOW 
DULL 
VERY DULL 
INOl 
20 
AL S 
■GiP 
\ 
AN 
U 
— - 
0 
-1°- 
— 
Ou 
ICn INTELLIGENT 
IMTELLICENT 
SLOW 
NTELLIOENT 
SLOW 
SLOW 
DULL 
VERY DULL 
Fig. 4. 
the normal curves for the school-boys' and school-girls' intelligence distributions 
are drawn, the total range of Intelligent -|- Slow Intelligent is taken as unit and 
the boys' and giils' distributions placed under each other on this basis. It will be 
seen from this standpoint that the girls' mean is slightly in excess of the boys', 
and if the girls' curve be now placed on top of the boys' (dotted curve of upper 
