561 
SECOND REPORT -1832. 
Besides the influence of change of temperature on the mag¬ 
netic tension, we have likewise to consider the change induced 
in the bar itself as a pendulum, by which its angular inertia is 
increased or decreased. A change of 20 or 30 degrees of tem¬ 
perature will cause very sensible differences in this way; for 
which a correction is requisite, adapted to each particular 
magnet. 
In ail experiments with the horizontal needle, it is evident 
that we do not measure the whole magnetic intensity of the 
place of experiment; we obtain only one of its resolved por¬ 
tions, except the place be exactly on the magnetic equator : 
hence it is desirable to preserve the horizontal position perfect. 
The following ready method is resorted to by the author, for 
determining, at any time, whether the bar be accurately sus¬ 
pended in an horizontal direction.—Let the bar be held by a 
fine filament of suspension silk near the surface of a reflecting 
fluid, such as mercury, or, what answers equally well, near the 
surface of water which has a little indigo dissolved in it; an 
image of the bar will, of course, be seen by reflection—when 
the lines of the bar and its image are parallel, the bar is accu¬ 
rately balanced in the horizontal position: the parallelism of 
the lines is readily detected by the eye. 
The author here details the method employed by him for 
suspending the bar, and preserving the horizontal position, 
without changing the angular inertia of the mass. 
The bar is suspended by means of a fine loop of silk, fixed 
in a very small hole, drilled through a piece of brass imme¬ 
diately in the centre; so that the axis of suspension may at all 
times pass through the centre of gravity of the mass, and the 
point of magnetic neutrality. Previously to the bar being mag¬ 
netized, it is carefully balanced horizontally, after which two 
points are impressed on it, on each side of the centre, at mid¬ 
way between the centre and ends. Two small sliders of plati¬ 
num, of equal weight, are then carefully fitted on the bar, so 
as to be easily moved on it: these bear a very small proportion 
to the weight of the whole mass, and are at first placed with 
their centres immediately over the above-mentioned points : 
when the bar is rendered magnetic, the inclination is corrected 
for any place, by moving one of the sliders from the centre, and 
the other toward it, each by an equal quantity. The distance 
through which it is requisite to move the sliders, in order to cor¬ 
rect the inclination, may at all times he extremely small: hence, 
we may consider the angular inertia of the mass as being near¬ 
ly the same, since the errors, which would otherwise arise from 
this cause, are made to neutralize each other, or very nearly so. 
