various Properties of Terrestrial Magnetism. 85 



deflection was reduced only ^jth part by an increase of 

 temperature equal to 43°. It may here be remarked, that the 

 effect of temperature on the reciprocal influence of the de- 

 flector and needle, is in part compensated by a correspond- 

 ing change of the effect of the earth's magnetism on the lat- 

 ter. I am about to apply to my instrument a simple con- 

 trivance for limiting the distance to which the magnetic tubes 

 may be drawn out or pushed in, according to the tempera- 

 ture existing at the time; but its influence is so inconsider- 

 able, and corrections can be so easily applied to the results, 

 that it does not seem necessary to describe it. It is always 

 desirable to note the temperature at the time of observation, 

 and a small thermometer is affixed to the face of the box for 

 this purpose. 



I am aware that a vibrating needle or any moderately light 

 body in constant motion, is liable to be much affected by tem- 

 perature in consequence of the mechanical* action of the air, 

 in which currents are produced by the application of warmth ; 

 thus experiments on the oscillations of the needle, and even 

 of the pendulum in a small degree, are liable to more or less 

 uncertainty on this account, the anomalous and fluctuating re- 

 sults obtained from the former being often very remarkable. 



When the apparatus is not in use, the ends of the needle 

 are received into sockets attached to brass springs fixed to 

 the side of the box, and which may be pushed forward by 

 screws passing through it. The deflectors at the same time 

 should be turned round so as to be parallel to the needle, but 

 with its poles reversed. With this precaution, I have reason 

 to believe that the magnetic forces will, within a moderate 

 time, become sufficiently permanent for making satisfactory 

 comparisons relative to the magnetic intensity in different 

 parts of the world: at any rate, if there should be any diminu- 

 tion of these forces after the instrument has been completed 

 twelve months, it will, (protected as the magnets are,) I be- 

 lieve, be very inconsiderable, and its ratio continually dimi- 

 nishing!. 



* This source of error having been referred to by others, I think it pro- 

 per to say that I believe I was the first person to announce it, early in 

 1832. There is also a paper of mine on the subject since inserted in 

 the Loud, and Edinb. Phil. Mag. vol. i. p. 310. 



•j- In corroboration of this opinion, I have found that a horizontal mag- 

 netic needle, or rather bar, coated with sealing-wax, and suspended by un- 

 spun silk in a box, in my underground cellar, nearly three years ago, has 

 suffered no sensible change of intensity in the last twelve months; or at 

 least, the time in which it perforins a given number of vibrations docs 

 not differ so much as one second in 1434". The temperature there fluc- 

 tuates very little, and it was at 52" when these comparative results were 

 obtained. 



