motions. In the northern hemisphere it is north-east deflected to east
as it approaches the southern trades. By the same reasoning, coming from
the north before it readies the surface, it ought to be also a
north-east wind above the lower westerly currents. Now it is an observed
fact, that while in the latitude of New York, for instance, the lower
westerly winds are to the easterly, as 3 or 4 to 1, in the highest
regions of observed clouds, the ratio is much increased; and according
to our own observations in this place,[5] we have never seen the highest
cirrus clouds moving westward. How then is this continual interchange
kept up? Assuredly we cannot have a current from the poles without a
contrary current to the poles. If we go into the arctic circle, we again
find the westerly and northerly winds predominating. If the current from
the equator follows the surface, the westerly winds ought to be
south-west. If it be above the surface wind, then the surface wind is
the polar current, and ought to be north-east. Whereas, from the
testimony of all who have visited these regions, the prevailing winds
are north-west. How can this be?

Again, it is proved that the upper current near the equator is also from
the westward--as near due west as possible. Take the latitude of St
Vincent. The difference between the cosine of 13° and radius applied
to the circumference, is about 600 miles, which would give 25 miles per
hour to the eastward, in lat. 13°. But to do this, it is necessary to
transfer it suddenly from the equator; for by a slow motion the easterly
tendency would be lost. Give it 24 hours from the equator to lat. 13°,
without any loss of easterly tendency, and it comes to that latitude
with a velocity of 38 miles per hour to the northward, and only 25 to
the eastward; we have, therefore, a wind from south-west by south. Yet
it is known that in the tropics the highest visible clouds move from the
westward. But as no such case could occur as a transfer in twenty-four