The wind abating and veering to the south, on the first of March, we
steered west, in order to get farther from Mr Bouvet's track, which was
but a few degrees to the east of us, being at this time in the latitude
of 46° 44' S., longitude 33° 20' E., in which situation we found the
variation to be 23° 36' W. It is somewhat remarkable, that all the time
we had northerly winds, which were regular and constant for several
days, the weather was always thick and cloudy; but, as soon as they came
south of west, it cleared up, and was fine and pleasant. The barometer
began to rise several days before this change happened; but whether on
account of it, or our coming northward, cannot be determined.[14]

[Footnote 14: It may be worth while preserving here the remark made by
Mr Wales. When off, and in the neighbourhood of Georgia, the cold was
much less severe when the wind blew from the south, than when it came
from the north. He assigns no reason for it, and perhaps the
observations were too limited to place and time to justify any general
inferences. It may, however, be suggested, with little risk of error,
that the northerly wind would be most loaded with moisture, hence the
cloudy sort of weather noticed during its continuance; and that, on very
well-ascertained principles, moisture is a considerable source of
cold.--E.]

The wind remained not long at south before it veered round by the N.E.
to the N.W., blowing fresh and by squalls, attended, as before, with
rain and thick misty weather. We had some intervals of clear weather in
the afternoon of the 3d, when we found the variation to be 22° 26' W.;
latitude at this time 45° 8' S., longitude 30° 50' E. The following
night was very stormy, the wind blew from S.W. and in excessively heavy
squalls. At short intervals between the squalls the wind would fall