132 Biometric Constants of English Brcmi-weiglits 
In f;ict tlie mean values are </s 4'11, $s 4'39 ; and we may, we hold, safely conclude 
that the increased variability does not make a class difference, but is peculiar to 
measurements made under some diflBculties*. 
Lastly, we may turn to the variation in age at death : 
TABLE XL 
Variation in Duration of Life. 
Group 
Standard Deviation 
Coeflicient of Variation 
? 
6 
? 
(Jeneral Hospital Pupulation ... 
1211 
13-69 
24-59 
29-83 
iSoeiety of Friends t 
17-13 
18-73 
28-13 
30-80 
Peerage and Lauded Genti'y t ... 
16-84 
27-61 
Yorkshire Tombstones J 
12-97 
16-40 
18-97 
24-53 
Oxfordshire Tombstones! 
12-70 
14-67 
18-55 
21-68 
London Tombstones; ... 
1314 
14-06 
19-56 
21-22 
Thus the general hospital population, while more variable in duration of life 
than the groups taken from tombstone records, which were all of married adults, is 
less variable than the pedigree records of either the Society of Friends or the 
Peerage and Landed Gentry, which were siinply for persons over 2L We think, 
therefore, that the general hospital population, while sensibly differentiated from 
the other classes in actual duration of life, is not sensibly differentiated in the 
variability of that duration. 
We may sum up then the result of the last two sections as follows : 
A general hospital population is sensibly differentiated in its average 
anthropometric characters from other groups of the same population. But there 
appears to be no markedly significant difference in variation. It would seem, 
therefore, that observations made on a hospital population may with due precautions 
be used for problems in variation, but they should not be used for drawing 
infei-ences as to the type of the surrounding general population without extreme 
care. 
(5) On the Relative Magnitude and Variability of English Brain-iueights 
compared with those of other Races. 
Professor Raymond Pearl's reductions enable us to form a fairly extensive table 
of comparative brain-weights. We have added a few results for stature and skull 
capacity. In the bulk of cases these are dependent upon few individuals, and are far 
from reliable, so that they must be treated as merely suggestions until further data 
are available. 
* See on difference between stature and corpse length Pearson : Phil. Trans. Vol. 192 A, p. 191. 
t Pearson and Beeton : Biometrika, Vol. i. pp. 60 and 64. 
X Biometrika, Vol. ii. p. 488. 
