10 
A Study in Criminal Anthropometry 
form 20-6 % of the 927 recorded temperatures; we will raise the normal-minded 
temperatures to 1719 by multiplying that row by the factor 2-336 retaining, however, 
only unit frequencies. We find approximately : 
Table XXXV, modified to give 10 % weak-minded only. 
Tevij^enUure. 
Intelligence 
■i 
=o 
1 
CI 
Cs 
Cs 
OS 
OS 
1 
OS 
1 
?^ 
OS 
1 
Cs 
<2> 
CO 
OS 
1 
OS 
OS 
CO 
OS 
1—1 
CO 
OS 
CO 
OS 
4) 
Co 
OS 
98-5— 98-6 
CO' 
Co 
OS 
J. 
CO 
OS 
05 
OS 
1 
OS 
CO 
OS 
'SI 
OS 
(3S 
1 
OS 
Cs 
OS 
OS 
05 
OS 
OS 
OS 
1 
OS 
OS 
CO 
OS 
OS 
1 
OS 
OS 
ij 
1 
OS 
OS 
cs 
Totals 
Weak-minded 
Normal-minded 
2 
5 
4 
7 
1 
49 
4 
70 
10 
145 
12 
201 
24 
236 
25 
299 
34 
248 
31 
236 
24 
114 
15 
61 
8 
23 
2 
12 
1 
5 
2 
191 
1719 
Totals 
2 
5 
4 
7 
50 
74 
155 
213 
260 
324 
282 
267 
138 
76 
31 
14 
6 
2 
1910 
Here the correlation r (found by the bi-serial method) is for this 10 % of the 
weak-minded 
r=- -231 ± -030, 
whereas with the 20 % value of the weak-minded 
r = - -258 i -030. 
We conclude from this result that even halving the percentage of the feeble- 
minded will only alter the correlation by less than its probable error, and that 
accordingly for the purpose we have in hand the bi-serial values of the correlations 
based on 20 % of weak-mindedness will give sufficiently accurate results. 
The reader is now asked to examine the last six lines of Table C. He will 
observe that, while nowhere very important, all the correlations of mentality with 
age, temperature, pulse, respiration, height and weight are significant; that the most 
important is weight, and that the signs are in agreement with the conclusions 
already reached, i.e. those with less mentality, or the feeble-minded, have less 
weight, height and age*, but greater temperature and quicker pulse and respiration. 
Examining the second column, which gives for the same variates the partial correla- 
tions for constant age, we see that no very substantial changes are made in the 
crude values ; they all remain still significant, and this could have been anticipated 
by merely noting the small age and variate total correlations at the top of the 
first column. We conclude again that the physical and physiological characters 
dealt with are significantly related to feeble-mindedness, but the relationships are 
of such a low degree that they would be of absolutely no service for the purpose of 
* The extent to which this lesser age of the weak-minded criminal is due (a) to earlier average 
conviction, resulting from earlier crime or easier capture, (h) to more frequent transfer to other 
institutions,, or (r) to earlier death is of course not determinable from the present data. But it has to 
be borne in mind that fraudulent offenders convicted late in life belong to a different social class, and 
are rarely feeble-minded. This class has usually greater stature and weight. 
