specially useful to the sailor, as its
indications relative to the winds are much the most certain. But it is
not, _alone_, to be relied upon. This is well settled, although the
reasons for it have not been understood. Why it should rise sometimes
before storms, in opposition to the general rule--or fall at others
without rain--or rise occasionally during the heaviest gales, has been a
mystery, and impaired the confidence in its accuracy and usefulness even
of the class of philosophers of whom Sir George Harvey spoke, in the
sentence quoted in the introduction. But, as I have already intimated, it
is all very intelligible.

I have said that the barometer has no fair weather standard--the mean of
30 inches at the level of the sea being an _average_ of the _fair weather_
elevations and the _foul weather_ depressions. Its fair weather position,
it would seem, must be above the mean, therefore, and as much above as its
foul weather depressions are below. But this is not precisely true. Its
extreme fair weather range is 31 inches, and it rarely reaches that; while
its lowest storm range is down to 28, and is the most often reached of
the two. My barometer stands about 40 feet above ordinary high-water mark.
It is not a "wheel," but an open, "scale" barometer, and a perfectly good
one. Its most reliable fair weather standard is about 30-30/100 inches. It
is its _most common summer, set fair position_, but that position is often
at other and different elevations, at other periods of the year, during
fair weather. The reader must observe for his own locality, and satisfy
himself what the most common set fair position for the barometer is, at
the different periods of the year, where he resides. When he has
ascertained this, he may apply the following principles to illustrate its
exceptional action, and in judging of the future of the weather:

1st. _As to its rise before storms._--Supposing it to have been
stationary, at or about a set fair position, _for the period_, and for one
or two or more days, a very _gradual_ and _moderate_ rise is an indication
of continued fair weather; and a _sudden_ and _considerable rise_ is
indicative of a storm. If the sudden and considerable rise occurs in the
latter part of spring, summer, or early autumn, it indicates a storm of
the _first_ or _third classes_ described in Chapter X., if in winter, a
storm of the _first class_ only. If the elevation is _very_ sudden and
considerable, the storm w