Postscript. 
Ixv 
116. The observations of both the force magnetometers have been corrected for 
temperature, it is conceived, with a considerable approximation to accuracy ; but not 
wishing to dogmatize in the use of a new mode of determining the temperature co- 
efficient, I have, with Sir Thomas Brisbane's leave, printed in all cases the tem- 
peratures of the magnets. In this, as in some other cases, I have preferred giving 
what may seem at present too much, rather than any one should afterwards have 
reason to find that I had given too little. 
117. All the reductions have been made by my present assistants, Messrs 
Welsh and Hogg, and by myself. Each computation has been performed twice, 
and that generally by different individuals. 
Makerstoun, June 1846. 
Postscript. 
Value of the Scale Divisions of the Bifilar Magnetometer in parts of the whole 
Horizontal Force. 
118. A consideration of the theory of the bifilar magnetometer will shew that it 
is assumed that the suspending wires do not act at all by any elastic force ; that, in 
fact, the force opposing the magnetic force is the resolved portion of that due to the 
weight suspended endeavouring to gain its lowest point, and, therefore, that if u be 
any angle from the magnetic meridian to which the magnet is deflected, the corre- 
SI II u 
sponding torsion of the wires being v (No. 35.), then is a constant ratio. If 
the assumption fail, there will be every reason to doubt the accuracy of the coeffi- 
cient k, which depends on sin v and its difference. Any considerable error was not 
suspected ; but the method described in the note, pages 2 and 3, having been found 
to answer so well for the determination of the coefficient for the balance needle, there 
was little doubt but that it would succeed much better for that of the bifilar magnet. 
Experiments were accordingly made when the previous Introduction was nearly 
through the press. 
119. If the equation of equilibrium for the bifilar magnet when at right angles 
to the magnetic meridian be (No. 35.) 
m X =/ 
and if a magnet, whose moment is M, be placed in the magnetic meridian, with its 
centre m the continuation of the bifilar magnet when at right angles to the magnetic 
meridian, and at a distance r from its centre, the resulting angle of deflection being 
a % equal n scale divisions, the equation of equilibrium will be (see the note already 
referred to), 
m IX -1 1 — J cos At>=/ 
MAG. AND MET. OBS. 1843. r 
