THE INTENSITY OF NATURAL SELECTION 
IN MAN. 
SECOND PAPER. 
By E. C. SNOW, M.A., D.Sc. 
The present paper is a supplement to the memoir of the same title issued last 
year*. It is not proposed to give here any account of the work which has been 
done in the attempt to elicit information on the more difficult subject of the nature 
of selection in man, but only to publish the correlations and regressions obtained 
by using an alternative measure of environment, and by varying the periods in 
which the effects of a selective death-rate can be detected. 
The adoption of another method of correcting for environment implies not the 
slightest shaking of my confidence in the adequacy and validity of that employed in 
the first memoir, but the further work was entered upon because the importance of 
the subject renders the comparison of the results reached by the use of the various 
possible methods particularly desirable. The only criticism I have seen of the 
mode of measuring environment used in the earlier work is that by the Editors of 
the Journal of the Royal Statistical Society. Had there been a tithe of evidence 
supporting the view adumbrated in that criticism the memoir would have been 
practically valueless. Fortunately, however, no arguments whatever have been 
put forward in favour of the view held by the statistical critics, and very cogent 
facts against that view have already been given f. It is quite beside the point to 
show that the corrected standard deviation of the total mortality in the two 
periods considered is only 5°/ 0 or 6%. That standard deviation is in a number 
of cases appreciably of the same magnitude as the corresponding measure of 
dispersion in the earlier of the periods used, and, moreover, is many times its 
probable error. It matters not whether that standard deviation be 6°/ 0 or '06 o / o 
or 60°/ o of the mean value. 
* Drapers' Company Research Memoirs, "Studies in National Deterioration," No. VII. Dulau 
and Co. 1911. 
t Biometrika, Vol. vm. p. 456, 1912. 
