J. Blakeman 159 
by aid of the known values of the coefficients of variation Vp, V,„ (see Table II.), 
and the known correlations 7> i, and r„.^ (see Table XIII.)*. 
TABLE XXIV. 
Index: Diametral Product/ Brain-weight. 
Sex 
Mean 
Standard 
Deviation 
Coefficient of 
Variation 
Correlation 
with Age 
Eegression 
CoefMcient on Age 
6 
2-8492 
± -0106 
•1703 
± -0075 
5^9771 
± •2636 
+ -11.53 
+ -0615 
+ •0016 
+ -0008 
? 
2-7978 
±-011.3 
•1628 
± ^0080 
5-8189 
+ -2863 
+ •1150 
± ^0696 
+ -0014 
± -0008 
The mean value of the index is closely the same, 2-8 for both sexes. The 
difference of the variations is also within the difference due to random sampling. 
The influence of age is, within the limits of the pi'obable errors, the same for both 
sexes, and for the sixty years covered by the data would not alter the index by as 
much as -1 or by sensibly less than four per cent. We have thus the interesting 
result that while age tends to shrink diametral product and brain-weight its effect 
on their ratio is comparatively insignificant. It tends slightly to increase it, the 
shrinkage of the brain-weight being somewhat more than the shrinkage of the 
diameters. Turning to the regression coefficients in Table XIX., we see that the 
shrinkage of brain-weight in 60 years is 132 grs. on 1328 grs. average, or roughly 
10 per cent., and the shrinkage in diametral product 247 cm.'' on 3783 cm.-' average, 
or roughly G'o per cent. We have thus the source of the 4 per cent, change in the 
index. It is, therefore, not surprising that the index increases and does not shrink 
like the simple characters with increasing age. 
Supposing the index were absolutely constant, then equations (iii) p. 158 show 
us that it would only be a very inexact method of determining brain-weight from 
diametral product; for the constant terms are here 25 per cent, of the total probable 
brain-weights. 
(13) General Conclusions. 
While the writers of the present paper would admit that many of its conclusions 
are only tentative as the data are too sparse, they yet consider that they will be 
useful as suggesting lines for further research and as confirming and supplementing 
some of the results of Pearl's more elaborate study. They hold that: 
(i) The biometric constants determined fit in reasonably well with the 
biometric results hitherto obtained for other brain-weight material. 
* See footnote, however, to Table II. 
