290 
A Statistical Study in Cancel- Death-Bates 
factor closely associated with these. Eadio-activity and various electric phenomena 
have from time to time been accused of producing cancer. The increased use of 
high tension currents is an undoubted fact in modern city life, but a glance at the 
varying rates as applying to different occupations will suffice to show that the 
explanation does not lie here. Another marked feature of modern life is the 
increasing use of preserved foods. But again the death-rate as evidenced by the 
occupation groups seems to negative this as a probable cause. For instance the 
agricultural rate is higher than that for the military and police. 
The increasing pressure of competition with its attendant nervous strain 
appears to be the factor in the aetiology of both diseases that will account for the 
phenomena. It has long been recognised that nervous shock and worry may be a 
determining cause in diabetes. 
At once another disease becomes associated with this factor in modern life. 
Insanity, also on the increase, should show a positive correlation with cancer. 
Unfortunately figures of the insane rates for the various cities in the United 
States are not obtainable, so on the suggestion of Dr Troup I have correlated 
suicides as being the nearest approach to a measure of the prevalence of insanity. 
Doubtless suicides form some guide to the stress and strain of life in the various 
districts, but other factors have probably also an effect in determining the prevalence 
of suicide. Religious beliefs, and moral training, may have a marked influence. 
The correlation found for cancer death-rates and suicides for the 40 cities, for the 
five-year groups, is : 
p= -3962 ±-0899, 
and the 15 States for the year 1906 : 
p = -4412 + -1402, 
and for the Races according to birth-place of mother, 1900 : 
p = -6010 + -1362. 
Although these values are not high they are possibly significant ; and in the case 
of the races of the same order as that found to exist between the insane enume- 
rated in hospital and suicides for the 15 States which is : 
p = -6062 + -1101. 
In the case of correlations found for the States and Races, it must be remembered 
that much greater fluctuations may be expected owing to the values being calcu- 
lated on the returns for only one year, while in the case of the Cities it was possible 
to obtain a steadier result by averaging the returns for five years. 
From the special report on The Insane and Feeble Minded in U.S.A., published 
by the Census Office and dealing with the returns of 1904, I was able to obtain 
figures for the insane as they apply to the various States. 
Particular caution is required in dealing with these figures, and a few extracts 
from the introduction to the report may enable the reader to estimate better the 
