Oscillations of the Horizontal Needle. 15 
retaining the tension of magnets, by placing the dissimilar poles 
in contact, either with each other or with a mass of soft iron, or 
both, is certainly very objectionable. A variable magnetic ten- 
sion is by no means unlikely to ensue. If the force of two small 
bars be measured by the method of vibration, and be then joined 
at their dissimilar poles by masses of iron, and again separated, 
they will not unfrequently shew very great differences when again 
vibrated as before. It seems, therefore, safer not to give needles 
intended for experiments on terrestrial intensity a greater degree 
of tension than they can themselves retain ; and when laid aside 
for use, to place them with reversed poles out of contact with 
each other, and at the intervals of some inches apart ; at least I 
have found bars prepared as above stated (16), and treated in 
this way, remain very constant. 
26. The possibility of change in magnetic tension, renders it 
desirable that we should be in possession of some means of de- 
tecting such change when it occurs. The method occasionally re- 
sorted to, of returning again to a given point, and taking the oscil- 
lationsas before, is not free from objection and inconvenience; since 
we are ignorant of the changes which may occur in a magnet, in 
consequence of change of situation, nor are we sure that the di- 
rective force is a constant quantity in a given place. We ought, 
therefore, to be enabled to detect changes on our needles in any 
place, and after small intervals of time, if required. The follow- 
ing appears a very delicate method of examining the tension of a 
magnetic bar. 
27. It is well known, that if a needle be allowed to vibrate 
in a ring of copper, it will be more speedily brought to rest, 
than if allowed to vibrate in an open space, although no sen- 
sible differences seem to arise in the number of vibrations per- 
formed in a given time*. The effect seems to be in some direct 
ratio of the magnetic energy. ‘The influence of a ring of copper, 
* Annals of Philosophy, Vol. IX. N. S. 
