84 
STATISTICS: R. PEARL 
tality curves, it is obvious that these centenarian deaths must be centered 
at some point. The finding of as close an approximation as possible to the 
correct centering point has been attacked along two lines, with the results 
set forth in this paper. 
In the first place, actual deaths occurring in the year 1916 in the Registration 
Area of individuals reported to the Census Bureau as having died at age 100 
or more were studied. Detailed statistics giving the distribution of these 
deaths by years was very kindly furnished me by Dr. William H. Davis, 
Chief Statistician of the Vital Statistics Division, Bureau of the Census. I 
wish to record here my obligation to Doctor Davis for furnishing me with this 
tabulation. 
i 
\ 
\ 
\ 
i 
1 
h — 
/ 
90 
80 
70 
60 
20 
115 120 
AGE AT DEATH 
125 
130 
185 
FIG. 1. DISTRIBUTION OF THE DEPTHS IN THE REGISTRATION AREA IN 1916, 
AT AGE 100 OR MORE 
These figures are set forth in table 1, and graphically in figure 1. 
It is apparent at once from the table and diagram that there are gross errors 
in the records of ages in these centenarian deaths. Unquestionably, some 
people do live to very advanced ages^ but it is equally clear that they do not 
occur in such numbers, nor at such advanced ages as would be indicated by 
these figures, taken at their face values. 
Particularly are the figures for the colored population obviously incorrect. 
One notes the outstanding humps in the colored female line at the quinquennial 
years, 100, 105, 110, etc. These, of course, have no foundation in fact. In 
