the commencement or termination of storms, or the prolongation of Mr.
Redfield's storm paths.

_Localities._--The three principal localities of storms are as
follows:--I. The western portion of the basin of the North Atlantic; II.
The China Sea and Bay of Bengal; and III. The Indian Ocean, more
particularly in the neighbourhood of Mauritius. The first two have
already been marked out as areas for the three-hourly observations; to
the latter, the remark as to extra observations under the head of
Desiderata will apply.

_Margins._--Mr. Redfield has shown that on some occasions storms have
been preceded by an unusual pressure of the atmosphere; the barometer
has stood remarkably _high_, and it has hence been inferred that there
has existed _around_ the gale an accumulation of air forming a margin;
barometers placed under this margin indicating a much greater pressure
than the mean of the respective localities. With regard to the West
Indian and American hurricanes--any considerable increase of pressure,
especially within the space marked out to the eastward of the 50th
meridian, will demand immediate attention. Upon the barometer ranging
_very high_ within this space, three-hourly observations should be
immediately resorted to; and if possible, _hourly_ readings taken, and
this is the more important the nearer the vessel may be to the 50th
meridian. Each observation of the barometer should be accompanied by an
observation of the wind--its direction should be most carefully noted,
and the force estimated according to the scale in page 21, or by the
anemometer. It would be as well _at the time_ to project the barometric
readings in a curve even of a rough character, that the extent of fall
after the mercury had passed its maximum might be readily discernible by
the eye. A paper ruled in squares, the vertical lines representing the