522 FRETS, THE INDEX CEPHALICUS. 

Table 19. Coefficients of Correlation. 
M. F. 
L and B 0.413 0.416 
L and I — 0.34 — 0.525 
B and I 0.532 0.567 


TSCHEPOURKOWSKY (1905). finds interracially the correlation of 
skulllength and -breadth nagative (tab. 1, r = —.321). If we 
pass from one race to another, TSCHEPOURKOWSKY remarks, we 
find that with the growth of length there is a decrease of breadth 
and vice versa. PEARSON (1903, p. 355) finds for maximum length 
and -breadth of: living head the interracial results M, = 184.64 
M3 = 140.0 and r;g = —.2917. He concludes that a principle of 
compensation applies. A longheaded race is likely to be a narrow- 
headed race and a broadheaded race a shortheaded race. This 
conclusion, he adds, is essentially interracial, not intraracial. The 
longheaded individual within the race, PEARSON continues, is likeiy 
also to be relatively broad, i. e. individuals within the race 
tend to have generally big or generally small heads. The principle 
of compensation only applies when we compare the mean lengths 
and -breadths of different races. 
As I have shown (p. 511) intraracially the principle of compen- 
sation applies too: grouping the headmeasurements according to 
lengthclasses, resp. breadthclasses we find that a long-headed 
length-class is relatively narrow-headed and vice versa. So with 
this restriction it applies that a longheaded individual is broad- 
headed too: there is a positive correlation between headlength 
and headbreadth, but for different lengthclasses there is a nega- 
tive correlation between headlength and headindex and a positive 
correlation between headbreadth and index. 
There is a slight negative correlation between stature and cranial 
index (SCHUSTER, 1911 finds r = —-.13; see also Craig 1911, 
PEARSON 1911, p. 68 note). 
As for the correlation of capacity and index (capacity and L, 
see MAcDONELL 1904) LEE (1901, p. 231) finds for skulls of the 
Aino race, that is dolichocephalic, a negative correlation between 

