'The Transactions of the Bombay Geographical Society. From January 1863 to December 1864. (Edited by the Secretary.) Volume XVII.' [‎196] (367/524)

196
EFFECT OF THE ATMOSPHERE UPON THE
compared with the diurnal variation of the temperature of the ah
within or without the observatory, an apparent relation would he found
to exist (just the opposite of what would be found in Europe) and a co
efficient might be deduced which would be so much the smaller, the
larger the temperature variation, so that the co-efficient for the greater
variations of external temperature would be less than that for the small
variations af internal temperature as Captain Montriou found.* But
that there should have been any tendency to similarity in the result
from the three tables was quite accidental, and due in that instance to
the previous erroneous correction for the temperature of the magnet.f
An evident method of determining whether the temperature of the
external atmosphere has any effect upon the bifilar readings indepen
dently of its effect in heating the magnet, is to heat artificially the room
containing the magnetometer, so that when the temperature of the
external air is falling, that of the magnet may be rising, since an increase
of temperature of the magnet causes a diminution of the scale reading.
If the increasing temperature of the external air caused an increase of
scale reading (as Captain Montriou supposed), the reading would change
by the difference of the two effects, but if the temperature of the external
air diminished, while that of the magnet increased, the change (accord
ing to the same supposition) would be equivalent to the sum of the
two effects. Such comparisons by artificial heating were made by mein
1843, and the conclusion was, that the effect on the scale reading was
due wholly to the temperature of the magnet, and was quite the same
whether the external temperature increased or diminished.^
There was still another assumption in connection with the result from
the Bombay Observations which merits notice. The hourly variations of
temperature of the circumambient air were assumed to be shown by a
* The same remark applies to the result derived from the third table of diurnal
ranges.
t In subsequent volumes of Observations it is merely stated that this correction
has been verified from subsequent years’ observations, but it may be safely asserted
that the agreement of results from the three tables could not be the same whatever
temperature co-efficient was employed.
t Makerstoun Observations. Transactions Royal Society Edinburgh, Vol.
III. page xlviii. Any difference must have been quite of a second order and not
appreciable in the discussions.