Vol. 6, 1920 
STATISTICS: R. PEARL 
429 
tion chosen as a standard of reference the actual values of 0 will practically 
always be negative for cities. The smaller these negative values are 
numerically the greater will be the proportion of older persons in the 
population concerned. In short this function 0 tells us not only the de- 
gree to which a given population deviates in its age distribution from a 
fixed standard age distribution, but also the nature of this deviation, 
whether on the one hand in the direction of a relative excess of aged, or 
9S 
AGE 
Fig. 2. Diagram showing age distribution of population, grouped in six age classes. 
The solid line gives the stationary life-table population, and the broken line the 
population of Atlanta. Total area under each curve = 100%. 
on the other hand in the direction of a relative excess of the young. 
Theoretically it is possible for two populations differing from one another 
in a compensatory way to give the same values for the index 0. But 
two populations which differ in age distribution in any fundamental re- 
spect v/hich could affect appreciably crude death rates will, in all popula- 
tions I have been able to test, give different values of </>, provided the age 
classification from which the function is calculated is finely enough divided. 
