y of the rate has been
_persistently downward_ from 26.7 per 1,000 in 1871 to 20.6 in
1886 and 17.4 in 1904. Commencing with the rate for the year
1871, the general death-rate of the white population of
Southern cities shows an _upward direction_ at different times
_during twelve years_, and a _downward_ direction _during
twenty-one years_, following in this respect practically the
same course as the corresponding death-rate for Northern and
Western cities combined. The year of _maximum mortality_ was
_1878_, due to a yellow fever epidemic, while the year of
_minimum mortality_ was, as in the case of the Northern and
Western cities, _1903_.

In reference to the table for the Negro population he says,[19]

Without exception, the death-rates are materially in excess of
the corresponding death-rates of the white population, but there
has also been in this case a _persistent decline_ in the general
death-rate from 38.1 per 1,000 in 1871 to 32.9 in 1886 and 28.1
in 1904. Commencing with the rate for the year 1871, the general
death-rate of the colored population of Southern cities at
different times assumed an _upward_ direction _during fifteen
years_ and a _downward_ direction _during eighteen years_,
departing in this respect from the corresponding mortality of
the white population of Southern cities and the general
population of Northern and Western cities, the tendency of which
was more distinctly towards a definite improvement. The year of
_maximum mortality_ for the colored population was _1873_, while
the year of _minimum mortality_ was _1903_.

The general correspondence and few divergencies of the two death-rates
are more clearly seen from the following diagram,[20] adapted from
Hoffman's study already cited:

[Illustration: Diagram II:
THE GENERAL DEATH RATE OF AMERICAN CITIES 1871-1904]

Other data[21] for two of the cities investigated by Mr. Hoffman, and
for three other cities (Atlanta, Ga., Charleston, S.C., and Richmond,
Va.) from 1882 to 1905 furnish results similar to his and indicate
likewise that while the general death-rate for the Negro population is
uniformly in excess of that of the white, _there is a tendency
downward_. For example, in Atlanta, Ga., the death-rates from 1882 to
1885 were for the white population, 18.22 per 1,000, Negro, 37.96;
from 1886 to 1890, white, 19.25, Negro, 33.41;