d by the rarefaction it causes draws
towards its region the land air, grown cooler, more dense, and heavier,
which continues thus to flow back till the earth, by a renovation of its
heat in the morning, once more obtains the ascendancy. Such is the
general rule, conformable with experience, and founded, as it seems to
me, in the laws of motion and the nature of things. The following
observations will serve to corroborate what I have advanced, and to throw
additional light on the subject for the information and guidance of any
future investigator.

The periodical winds which are supposed to blow during six months from
the north-west and as many from the south-east rarely observe this
regularity, except in the very heart of the monsoon; inclining, almost at
all times, several points to seaward, and not unfrequently blowing from
the south-west or in a line perpendicular to the coast. This must be
attributed to the influence of that principle which causes the land and
sea winds proving on these occasions more powerful than the principle of
the periodical winds; which two seem here to act at right angles with
each other; and as the influence of either is prevalent the winds draw
towards a course perpendicular to or parallel with the line of the coast.
Excepting when a squall or other sudden alteration of weather, to which
these climates are particularly liable, produces an irregularity, the
tendency of the land-wind at night has almost ever a correspondence with
the sea-wind of the preceding or following day; not blowing in a
direction immediately opposite to it (which would be the case if the
former were, as some writers have supposed, merely the effect of the
accumulation and redundance of the latter, without any positive cause)
but forming an equal and contiguous angle, of which the coast is the
common side. Thus, if the coast be conceived to run north and south, the
same influence, or combination of influences, which produces a sea-wind
at north-west produces a land-wind at north-east; or adapting the case to
Sumatra, which lies north-west and south-east, a sea-wind at south is
preceded or followed by a land-wind at east. This remark must not be
taken in too strict a sense, but only as the result of general
observation. If the land-wind, in the course of the night, should draw
round from east to north it would be looked upon as an infallible
prognostic of a west or north-west wind the next day. On this principle
it is th