58 
Variation and Correlation in Brain-Weight 
the female. It has been shown above that there is in general no significant 
difference between the sexes in respect to variability in bi'ain-weight, so that 
the usual formula of " low variation and higli correlation " connoting stability of 
type does not hold in this case. It is possible that we have here an expression 
in a particular case of a greater general " evenness " in the females of the bodily 
changes accompanying increasing age, which in turn might be due to the generally 
more even environmental conditions to which women are subjected. It is note- 
worthy in this connection that the correlation in respect to duration of life is 
generally higher between pairs of female relatives than between pairs of male 
relatives *. 
The linearity of the regression of brain-weight on age is of interest as possibly 
indicating a fundamental difference in the modes of action of the biological 
processes of growth on the one hand and senescence on the other. When growth 
in absolute magnitude of a character is plotted on a base line of age the result 
is usually a curved line (cf for example the 12 — 25 portion of Powys' stature 
curve, loc. cit.), which implies, of course, that the amount of increment in the 
character in question is not the same for each unit of time. On the other hand 
in the case of decrease of brain-weight and stature with advancing age the 
decrement seems to be practically uniform for each unit of time. If these relations 
should prove to be generally true they would furnish a very interesting field for 
further study and analysis. 
Turning now to the relation of brain-weight to stature, we have in Table XV 
the values of r, rj and 2jV) for all the brain-weight and stature correlations. 
TABLE XV. 
Linearity of Regression. Brain-weight and Stature. 
Eace and Series 
r 
2m 
? 
? 
$ 
? 
Swedish (Total) ... 
•1830+ -0320 
•3490+ ^0388 
■2439 
•3847 
■1612 0- 
•I6I80- 
„ (Young)... 
•1796+ -0403 
•3390 + -0530 
•2837 
•4738 
•2196 0- 
•3310 0- 
Hessian (Total) ... 
•1823+^0299 
•1828+^0389 
•2864 
•3215 
•2209 0- 
•2645 0- 
(Young)... 
•1741 + -0383 
•1809+^0496 
•2714 
•3807 
•2082 0- 
•3350 o- 
Bavarian (Total) ... 
•1664+ -0343 
•2236+ -0413 
•2262 
■3270 
•1532 o- 
■2386 0- 
Bohemian (Young) ... 
•2034+ -0397 
•2168+ ^0557 
•2419 
•3591 
•1309 0- 
■2863 0- 
Again, the values of t] and are seen to diverge considerably from r and 0 
respectively. It is clear that on the whole the regression of brain-weight on 
stature approaches less closely to linearity than does the regression of brain- 
weight on age. This is, I think, a somewhat remarkable result, and one not 
likely to have been foreseen. 
* Beeton and Pearson : Biometriha, Vol. i. p. 60. 
