12 
A Studi/ in Criminal Antliropometry 
diagnosis, or for discriminating the weak-minded from the normal-minded. An 
examination of the Tables XXXVII, XXXVIII and XLII, XLIII shows that the 
range of the normal in every case is practically coextensive with, or more extensive 
than, that of the feeble-minded. This is in good accordance with the results already 
reached by Jaederholm and Pearson from other data : see Questions of the Day and 
Fray, viii. "The Continuity of Mental Defect," and ix. "On the Graduated 
Character of Mental Defect." With regard to the further partial correlations, it 
will be seen that the relationships between the physiological variates (temperature, 
pulse and respiration) and mentality are not much modified by making the physical 
variates constant. The relationship between mentality and temperature reaches its 
maximum for constant weight and height, but the gain is only about its probable 
error; that for mentality and pulse is more considerable being three times the 
probable error for constant age and weight ; while that for mentality and respiration 
is a maximum for constant age and height, but the increase is less than half the 
probable error. Finally it may be noted that the relationship of mentality to 
weight is somewhat intensified if the temperature and pulse rates be made 
constant. 
In no case, however, do we reach correlations of the slightest importance for 
diagnostic purposes. 
(4) On the Inter physiological Correlations. 
The next point to which we must turn our attention is the relationship between 
temperature, pulse and respiration. We have the following series of correlations 
from Table C. 
Coirelated Pair 
Crude 
Correlations 
Pai'tial 
Correlations for 
Constant Age 
and Weight 
Partial 
Correlations for 
Constant Age 
and Height 
Temperature {T) and Pulse (P) 
Temperature (T) and Respiration (i?) 
Pulse (P) and Respiration (7?) 
+ •288 ±-020 
+ ^142±^022 
+ .060±-022 
+ •308 ±-020 
+ ^163±^022 
+ •081 ±-022 
+ ^314±^020 
+ ^158±-022 
+ •059 ±-022 
All these correlations are significant, but the sole one of any importance, and 
even that is of little value, is the relation between temperature and pulse. The only 
comparative material is that of Korosy reduced by Julia Bell (Biometrilia, Vol. viii. 
p. 232). She found for pulse and respiration -|- -lOSi -042, in excellent agreement 
with our present value. Her values for pulse and height, — •072 ± -042, and for 
respiration and height, — -042 ± -042, are also in accordance with ours considering 
their probable errors. The smallness of the relation between pulse and respiration is 
noteworthy, and none of these relations are substantially modified if we make age 
and weight or age and height constant. In fact temperature, as an examination 
of Table C shows, has no relation to either height or weight, while greater stature 
