23 
the intensity of mag?ietic forces ^ &c. 
After having made the preceding observations, and con- 
cluded those on the diurnal changes in the points of equi- 
librium, I proposed applying the balance of torsion, as another 
means of determining the variations in the magnetic inten- 
sities, arising from changes in the temperatures of the mag- 
nets ; but in my first observations with this instrument, there 
were such discrepancies, arising from the over torsion of the 
wire in consequence of its want of elasticity, being silver and 
too fine for the weight of the needle, and likewise too short, 
that, although they pointed out very clearly the same general 
results which I afterwards obtained from unexceptionable 
observations, I would not make use of them for determining 
p 
a mean value of A . to be applied to the correction of the 
observations on the diurnal changes, for the variation in the 
temperature of the magnets. As a considerable time inter- 
vened before I had an opportunity of repeating these experi- 
ments, and in making them, I had, by increasing the tem- 
perature of the magnets beyond a certain point, permanently 
destroyed a portion of their intensities, I considered it better 
to obtain the mean value of which was requisite, from 
the results of the experiments made more nearly at the 
same time as the observations which it was my object to 
reduce. I therefore determined this value from the preceding 
results. 
The temperatures of the magnets in the observations on 
the daily changes of the points of equilibrium were, with a 
few exceptions above and below, comprised between 54° and 
p 
65®, and I determined the mean value of A^ to be applied 
to the correction of these observations in this manner : from 
