correlation is made with the mortality at each age class:

TABLE II

Under 1 year 0.3647 +- 0.0986
1-19 years 0.4884 +- 0.0816
20-39 years 0.6228 +- 0.0656
40-59 years 0.5028 +- 0.0801
60 years and over 0.2577 +- 0.1001

"A peculiar configuration is observed in these coefficients in that a
quite pronounced positive correlation exists at the central age
group, but disappears with some regularity towards both extremities
of life. If the mortality has any influence upon the natality this
cannot be in the form of replacement of lost infants and deceased old
people, therefore, as has frequently been suggested. That a high
death-rate at the child-bearing age should be conducive to increased
fertility is absurd, neither does it seem likely that a large number
of children should make the parents more liable to diseases which are
prevalent at this period of life. The reasons must, then, be looked
for in a common factor.

"Now the only disease of importance representing the same age-curve as
do the correlation coefficients is tuberculosis. This disease causes in
France 2 per cent. of the deaths under one year, 24 per cent. of the
deaths from 1 to 19 years of age, not less than 45 per cent. from 20 to
39, 18 per cent. at ages 40 to 59, and less than 2 per cent. at the
ages over 60. Will a high tuberculosis mortality, then, be conducive to
great fertility, or do we have to fear that a decrease of the natality
will be the result of energetic measures against tuberculosis? Hardly.