Shinkishi Hatai 431 
and Macdonell, 02, Pearson’s conclusion, so far as the measurements of 
the skulls are concerned, is not well supported. For instance the total 
averages of the coefficients of variation in man (Aino and Naquada races 
are excluded) are 4.99 per cent while in woman it is 4.84 per cent. 
Even if we include Aino and Naquada, the relative variability is still in 
favor of man. The same is true also for the albino rat (4.31 per cent 
male, and 4.17 per cent female). Therefore so far as the data at hand are 
concerned the several measurements of the male skulls show a general 
trend to a greater variability than those of the female. Since in all cases, 
the number of the skulls examined is not large, it is evident that this 
point needs still further study. 
DETERMINATION OF THE COEFFICIENTS OF CORRELATION. 
The degree of correlation between any two characters is usually deter- 
: e(2. y.) Le 
mined by the formula: Y = a ee where x, y are deviations from the 
) aa. 
means of the two correlated characters and 6,, 8, the respective standard 
deviations. It will be advantageous to discuss the coefficients of cor- 
relation under five headings. 
(a) Correlation between the length of the entire skull and the other 
cranial measurements. 
TABLE IV. 
Male. Female. Difference. 
Length of entire skull and zygomatic 
SW al Cli Fee or rete rato ova) ane sucetne aeveh ears wens; a= Adplal .836 = .029 113 + .031 
Length of entire skull and fronto- 
OCGipitalelensth Waa. asde ce ciate .946 + .010 .956 + .008 —.010 + .013 
Length of entire skull and length of 
MASAI DOM C2 vschowereke) «on excel eclereterts .845 + .027 .890 + .020 .044 + .034 
Length of entire skull and squa- 
MOSAlCISCANGE! cic aceiie elastance 582 + .061 -309 == .085 20 == L065 
Length of entire skull and height ; 
Olive SUM cS his ce occ teicle ste oie iciotes -555 + .064 314 + .085 ALOR 
As one would expect, the degree of correlation in the first three cases 
is very high in both sexes. Since the length of the skull depends on the 
length of the nasal bone and the fronto-occipital length, any change in 
the length in the entire skull must involve a change in the length of either 
the nasal bone or the fronto-occipital length or in both. The correlation 
shows, however, that change in the length of the skull is associated with 
changes in both the nasal bone and the fronto-occipital length; the latter 
