40 Mr. Christie on the effects of temperature on 
room as those from the 20th to the 27th of April, described 
in my former paper. The compass was placed on an hori- 
zontal table, with its centre at the distance of 5 feet from 
the middle of the only window in the room, and which was 
nearly in the direction of the magnetic meridian from it. I 
mention this circumstance, not that I myself consider it of 
importance, but as a datum for those who may be disposed 
to attribute the diurnal changes in the direction of the needle 
to the influence of light. The only iron in the room is a 
large lock to the door and the weights to the window, which, 
when the observations were made, were always in the same 
position. The magnets were placed on the table with their 
axes, as nearly as I could adjust them, in the meridian, the 
north pole of each being, as I have before mentioned, to- 
wards the north, and the distances of their centres from the 
centre of the needle 21.21 inches.* 
The method which I at first adopted for determining the 
changes that took place in the temperature of the magnets, 
was by placing a thermometer with the bulb near the south- 
ern extremity of the north magnet. In this manner I conti- 
nued to observe for five days ; I then placed the bulb of the 
thermometer o?i the southern extremity of the north magnet ; 
and continued the observations for five days longer. I con- 
sider that the changes in the thermometer would, in either 
case, very nearly indicate the changes in the temperature of 
the magnets, especially as those changes were very gradual, 
and did not exceed 10° during the whole time in which the 
observations were made. 
In the present state of our knowledge respecting the causes 
of magnetical phaenomena, it is difficult to say how far 
