chronometers by the proximity of masses of iron. 363 
greater; but it would still depend upon the original adjust- 
ment of the machine, whether the result would be to accele- 
rate or to retard its action ; that is to say, it would depend 
upon the contingency, whether the chronometer had a ten- 
dency to gain, or lose, in short arcs, which I am informed is 
nearly an equal chance, if it proceed from the hands of a 
scientific workman ; but that, in general cases, the proba- 
bility is, that the watch will lose in large arcs, and gain in 
small ones. 
The experiments and observations which Mr. Fisher de- 
scribes as having been made with a strong bar magnet, 
brought within two inches of the balance, I consider to be 
perfectly distinct in their nature from those which were made 
by him on board and on shore at Spitzbergen ; for a magnet 
of such power, brought within the distance of two inches of 
any small piece of steel, will, whether the latter be previously 
magnetic or not, impress upon it a strong temporary derange- 
ment of its latent magnetism, and give to the part nearest 
the magnet, a contrary pole to that by which it is opposed ; 
and consequently, there will exist between the balance and 
the magnet a strong power of attraction sufficient to cause 
that acceleration so strongly indicated in Mr. Fisher’s ex- 
periments; and this will be the case whichever end of the 
magnet is opposed to the balance, and to whatever part of 
the latter the application is made ; because, in this instance, 
the effect does not depend upon the previous magnetic state 
of the balance, but upon that temporary state excited by the 
proximity of the magnetic bar, and which ceases when the 
bar is removed. 
This explanation will not, I conceive, apply to the action of 
MDCCCXXI. 3 A 
