M. A. Lewenz and K. Pearson 
387 
We have in this formula the mean value equal to that of the observed races, 
but the mean error is only reduced '2, in other words our G. F. is not substantially 
bettered, it is as good a fit as can be got for this series from a linear formula. 
But this best fitting line is interesting from another point of view, which is 
exhibited in the accompanying diagram. We have plotted here the observed 
capacities of 19 races, the above 10 races and the 9 doubtfully measured races from 
Barnard Davis; the horizontal abscissa is the product in cubic centimetres of the 
three arcs U, S and Q. We have added the cephalic indices. The straight line A A 
is the inter-racial regression line based on the 10 races. We see that with the 
If" 
lie . 
t 
B 

ro— 
r^) 
rr) 
=) 
(rr) 
/ 
/ 
/ 
/ 
/ 
/ 
-/ 0* 
/ 
/ 
(ro) 
-si 
/ 
exceptions of the Malays and the Marquesans*, both fairly close to the inter-racial 
line A A, all the other races above the line are modern Europeans. In other words 
it would appear that the modern European races gain their advantage in skull 
capacity over the lower or more primitive races largely by thinner bone. Of the races 
* "Without exception the finest race of people in this sea. For fine shape and regular features 
they perhaps surpass all other nations." Cook, Voyage towards the Soutli Pole, 1777, Vol. i. p. 308. 
