character; and, I have reason to believe, from observation, in one
or two cases, that where belts of rains and showers begin, over _any
locality_ in the United States, they may assume this character. I have
been in Saratoga when an easterly storm commenced _south of that place_;
the condensation and mackerel sky being visible at the south, and no cloud
formation or rain occurring there at the time, and have traced it
afterward as a belt which had a lateral extension south-eastward. Leaving
that place immediately after a belt had passed south, I have overtaken it
by railroad, and run into it again before arriving at New York; and
witnessed its subsequent extension south-eastwardly, out over the
Atlantic. I have witnessed the approach of such a belt in the spring, at
Sandusky, upon Lake Erie, and its passage over to the S. E., followed by
the N. W. wind, as Mr. Bassnett describes them at Ottawa, and run under
the attenuated edge of the same belt, on the same day, on the way to
Pittsburg, leaving the N. W. wind behind, but finding it present again
with clear sky on the following morning. I have seen hundreds of them
approach from the north, and pass to S. E., out over the Atlantic;
followed by the N. W. wind in spring and autumn. This class of storms pass
off toward, and doubtless over the track, of our European steamers and
packets. I know this, for I witness it nearly every month in the year. It
is not a matter of speculation, but of actual, long-continued observation.
Probably, as one approaches the Gulf Stream, and when over it, its induced
winds may be more violent. It is time our navigators understood this; and
that all the gales of the North Atlantic, certainly, are not rotary; and
do not approach from the S. W. in the same manner as the class
investigated by Mr. Redfield do. Where a fresh southerly or south-westerly
wind is followed by any considerable cirro-stratus or stratus-condensation,
it is usually of this character.

The following diagram exhibits the peculiarities of this class of storms.
It is intended to represent the same storm or belt of showers, on _two
successive_ days, and, of course, its usual rate of southerly extension:

[Illustration: Fig. 21.]

This class of storms, or belts of showers, present the following
succession of phenomena in summer:

1. Still warm weather, one or more days.

2. Fresh southerly wind, one or more days; if more than one, dying away at
the S. W., at night-fall, bu