Raymond Pearl 
65 
conditions preceding death. Yet as Donaldson* has pointed out if we could get 
this character free of disturbing ante-)nortem effects it would furnish the best 
measure of the general " size " of the body. Since " size " of the body is one of the 
things to which brain-weight should first of all be related T decided to include in 
this study a consideration of the correlation of brain-weight with body-weight. 
I was fully aware of the weakness of the material and expected much worse results 
than those actually obtained. A glance at Table I will recall the facts as to the 
distribution of the frequency of body-weight. The means are of course lower than 
they would be for the living population of the same race and age. This is to be 
expected. The variation is large but there is very good accord between males and 
females in this respect. Turning now to the correlations I have arranged in 
Table XVI the values of q and Sj^/ for the correlation between bram-weight and 
body-weight. 
TABLE XVI. 
Linearity of Regression. Brain-weight on Body-weight. 
Race and Scries 
r 
V 
2 
6 
? 
? 
6 
9 
Bavarian (Total) 
•1671 ± -0343 
•2260 ±-0412 
•1845 
•3189 
•0782 0- 
•2250 o- 
The result for the males is very satisfactory, indicating a quite reasonable 
approach to linearity of regression. The female series does not give so good a 
result, but it will be recalled that throughout the work this Bavarian female series 
has been found to be somewhat abnormal. The regression lines are shown in 
Diagrams XVIII and XIX. 
DiAGEAM XVIIl. Probable Brain-weight for given Body-weight. 
Bavarian d Total. 
( 
) 
( 
) 
( 
^ — 
< 
) 
( 
( 
) 
( 
( 
I ^ 
) 
' ( 
> 
< 
■> 
( 
) 
g 1450 
S 1300 
S 
=5 
27-5 32 5 37 5 42 5 47-5 52 5 57-5 62 5 67-5 72-5 77 5 82 5 87 5 92-5 
Body-wcirjht ( = TI',,) in kilograms. 
H. H. Donaldson : The Groicth of the Brain. 
Biometrika iv 
