If we observe the fluctuation of the thermometer by the side of the
mortality of the nation at large, no calculable relationship seems, at
first sight, to be traceable between the one and the other. But if, in
connection with the mortality, care be taken to isolate cases, and to
divide them into groups according to the ages of those who die, a
singular and significant series of facts follow, which show that after
a given age a sudden decline of the temperature influences mortality
by what may be considered a definite law. The law is, that variations
of temperature exert no marked influence on the mortality of the
population under the age of thirty years; but after the age of thirty
is reached, a fall of temperature, sufficient to cause an increased
number of deaths, acts in a regular manner, as it may be said, in
waves or lines of intensity, according to the ages of the people. If
we make these lines nine years long, we discover that they double in
effect at each successive point. Thus, if the, fall in the temperature
be sufficient to increase the mortality at the rate of one person of
the age of thirty, the increase will run as follows: 1 death at 30
years of age will become 2 deaths at 39 years of age, 4 at 48 years, 8
at 57 years, 16 at 66 years, 33 at 75 years, and 64 at 84 years.

In these calculations nothing seems to be wanting that should render
them trustworthy; they resulted from inquiries conducted on the
largest scale; they were computed by one of our greatest authorities
in vital statistics, the late Dr. William Farr, and they accord with
what we gather from common daily observation. They supply, in a word,
the scientific details and refinements of a rough estimate founded on
universal experience, and they lead us to think very gravely on many
subjects which may not have occurred to us before, and which are as
curious as they are important.