56 
Inheritance of the Duration of Life 
of birth as far as longevity is concerned. Out of 1606 pairs of sisters the elder 
lives hunger in 922 5 cases ; out of 2247 pairs of brothers the elder lives longer in 
1303'5 cases. Or, in 57"4 per cent, of cases the elder sister, in 58"0 per cent, of 
cases the elder brother lives longer. Summing up our results as for the case of 
adults we have for all brethren of whatever age at death : 
Brothers 
Sisters 
Moan excess in life of elder . . . 
Standard deviation of excess... 
Mean interval between birtlis 
Correlation 
Number of pairs 
4-612 yrs. 
30-2209 yrs. 
6-717 yrs. 
•1096+ -0141 
1606 
3-232 yrs. 
29-7849 yrs. 
6-8400 yrs. 
■1352 + -0165 
2247 
Now it will be seen that there is no sensible change in the correlation whether 
we take all pairs of brethren or only adult pairs, nor is there much difference 
between the mean intervals between births. But the elder brother has slightly 
gained and the elder sister distinctly lost in excess of life by taking brethren of all 
ages instead of merely adult pairs. I attribute this to the much larger correlation 
between the ages at death of infant sisters than we find in the case of infant 
brothers (see p. 60 below). 
The following formulae enable us to predict the probable excess of life (e) of an 
elder brother or sister from a knowledge of the birth interval {i), both being taken 
in years : 
Elder brother: e = -72821 - -2795. 
Elder sister: e = •8525{ - 2-5991. 
Thus a brother born ten years before another brother has probably seven years' 
greater duration of life ; a sister born ten years before another sister has probably 
about six years' greater duration of life. Theoretically the younger brother has the 
advantage of the elder until i = ■2795/-7282 = -38 years about — an impossible 
difference. Theoretically the younger sister has the advantage of the elder 
sister until i = 2-5991 /■8525 = 3-05 years about, or until the third year say of inter- 
val of birth. These results are apparent on Figs. 1 and 2, which give the regression 
lines, and are quite sufficient in practice for the determination of the probable 
difference in duration of life. 
For the present we have contented ourselves by showing the importance of 
seniority in birth for the appreciation of longevity. But it will be clear to the 
reader that the correlations between longevity and actual standing in a family and 
between longevity and the ages of the parents at the birth are matters of much 
interest and deserve full investigation. 
