86 
STATISTICS: R. PEARL 
2. Inasmuch as the hfe table values given in table 4 are certainly nearer 
the true values than those of table 2, they may well be adopted as the basis 
of centering deaths over 100 in biostatistical computations. 
3. Accordingly in this laboratory the deaths of whites, both male and 
female, occurring at 100 years and over will be assumed to center at 101.7 
TABLE 4 
Mean Age at Death and Standard Deviation in Age at Death of Centenarians, 
FROM United States Life Table Figures 
GROUP 
MEAN 
STANDARD 
DEVIATION 
101.66 
1.32 
101.73 
1.49 
101.70 
1.42 
102.32 
1.92 
years, and those of negroes at 102.0. The latter value is probably still too 
high, but unless one repudiates the statistical data entirely, which would 
amount to throwing away the child with the bath, there is no warrant for 
arbitrarily taking a lower value. 
1 Pearl, R., Biometrika, 4, 1906, (510-516). 
2 Cf. the authentic case reported by Rodriquez (in Siglo med., Madrid, 53, 1906, (341- 
343) > of death at 124 years of age. 
3 Clark, F. C, Providence, Med. J., 10, 19()9, (143-158). 
< Glover, J. W., United States Life Tables, 1910. Bureau of the Census 1916. 
