Madeline H. Whiting 
9 
Now it will be seen that while all these correlations are significant except age 
and height*, none of them are of any considerable magnitnde, or capable of 
producing any appreciable influence on the observed comparable differences. 
Further since pulse and respiration become very slightly quicker with age, they tend 
to emphasise and not to reduce the observed differences between these characters in 
weak-minded and normal-minded. Age might account for one-ninth of the observed 
difference in weight, or roughly for less than 1 lb. out of the nearly 8 lbs. difference 
observed; and, as before, age accounts for a very inconsiderable portion of the 
observed temperature difference. We are thus justified in concluding that the 
observed physical and physiological differences are quite independent of age 
diiJerencesf . We are able to confirm these results by another method of 
investigation, which well illustrates the insignificance of the age correction. In 
our opinion "weak-mindedness" must depend largely on the personal equation 
of the recorder^, and there is no definite bounding line between the weak- and 
the normal-minded. In other words we may consider mental state as an approxi- 
mately continuous variate, and one not widely divergent from the C4aussian in 
frequency distribution. Now it is known that in school-children the feeble-minded 
do not amount to 2 %§; and it is probable that in the community at large their 
number does not even reach this. On the other hand the criminal class is largely 
recruited from the less-intelligent section of the general population. Among 
inebriate and prostitute women probably 40 % to 60 % may be classed as feeble- 
minded ||, and though among male criminals such a high percentage is not reached, 
they must bulk very considerably. Goring in his invaluable Report^ considers 
the percentage to be 10 to 15. In our present data we have 100 weak-minded 
individuals and 400 normal-minded, or we have taken 20 % and calculated the 
correlations between intelligence and the physical and physiological characters 
on this basis. But it is of interest to observe what order of change would be made 
in our results by taking a lesser percentage. Consider, for example, Table XXXV 
for intelligence and temperature. Here the temperatures of the weak-minded 
* Of course the correlation of age and height is largely misleading, the younger criminals are still 
growing, the older are slowly losing height. Thus the regression is really skew. For example, the 
value of the correlation ratio •q for height on age is -147. If, however, we take the correlation of age 
and height for criminals of over 20 years, at which age other investigations (see Biomefrika, Vol. i. p. 47) 
show that height begins to diminish, the negative correlation is only —•01195 corresponding to merely 
a loss of -0022 inches per year of past prime life, and therefore quite negligible. 
I Except in the matter of age and weight Table B shows us that judged either by absolute or relative 
variability there is no practical difference in variability between weak-minded and normal-minded 
criminals. 
{ On the continuity of Weak-mindedness, see Questions of the Day and the Fray, No. viii. "The 
Continuity of Mental Defect," by Karl Pearson and Gustav A. Jaederholm; No. ix. "On the 
Graduated Character of Mental Defect," by Karl Pearson (Cambridge University Press). 
§ See A Preliminary Study of Alcoholism in Adults, by Amy Barrington and Karl Pearson (Cambridge 
University Press), p. 9. 
II A Second Study of Extreme Alcoholism, in Advlfs, by David Heron (Cambridge University Press), 
p. 15. 
*i See The Evglish Convict, a Statistical Study (Wyman and Sons), p. 254. 
