234 
Variation and Correlation of the Human Skull 
ranged from "74 to ".98*. A similarly high correlation has also been shown to 
exist between the first finger joints, the coefficients ranging from •82 to ■91'f". 
I will now examine in detail the results shown in the Table, and compare 
them with the results obtained for other races, incorporating in my tables the 
comparable material which C. D. Fawcett collected in her Tables XV., XVI. and 
XVII. 
Length, Breadth and Height Correlations. 
In the English series we observe that the feniales are much more highly cor- 
related than the males in all three pairs of characters. In this respect they agree 
with the German series, but here the comparison is unsatisfactory so far as height 
is concerned, the German height being auiicular, the English basio-bregmatic. 
English also agree with Naqada in so far as L and H is the most closely associated 
pair in both races. On the other hand, it is the males who are the more highly 
correlated throughout in Ainos and Naqadas, and also in French except for the 
pair B and H, where the sexes agree ; here again, however, it must be noted that 
in the case of the Ainos we are dealing with the auricular height. 
In absolute value of the correlation coefficients we observe that our series 
agrees fairly closely with the French in the pairs L and and B and H% and 
with the German in L and 5,/, but in no others. 
Correlation of Breadths. 
Little has hitherto been done in working out the correlation between cephalic {B), 
frontal {B') and bizygomatic {J) breadths: the following table gives all the material 
that I have been able to find : 
TABLE XII. 
Correlation of Head Breadtht 
Kace 
B an 
B and J 
B' and J 
c? 
? 
? 
? 
English 
•4.32 
•367 
•468 
•414 
•620 
•497 
Naqadas | 
•111 
•044 
•569 
Thebaii Mummies | 
•298 
•250 
•342 
Oraon Tribe t 
•587 
•409 
•554 
(living head) 
There is fairly close agreement between English skulls and living Oraons in all 
three correlation coefficients. The correlation between frontal and bizygomatic 
breadths is very much the same in Naqadas as in English and Oraons, and in all 
* Lee and Pearson: R. S. Proc. Vol. 61, pp. 349, 350. 
+ Whiteley and Pearson: R. S. Proc. Vol. 65, p. 130. See also Lewenz and Whiteley: Biometrika, 
Vol. I. pp. 345—60. 
t Biometrika, Vol. ii. pp. 349—351. 
