^11 



observation has been the veitical component of the intensity, 

 the instrument for the measurement of which has been al- 

 ready submitted to the notice of the Academy. The prin- 

 ciple of tliis instrument, it will be remembered, is to balance 

 the vertical component of the magnetic force by a fixed 

 weight, and to observe the changes of the position of equili- 

 brium, under the action of the changing force. Unexcep- 

 tionable as this principle is in theory, the accuracy of the 

 results has not been commensurate with that of the other 

 two instruments. This inferiority is to be traced to the 

 large influence which the unavoidable errors of workman- 

 ship must necessarily have on the position of equilibrium of 

 a magnet supported on a fixed axle. It has been shown that 

 the effect of magnetizing a bar, under the most advantageous 

 circumstances of form, and at the part of the globe where 

 the vertical component of the magnetic force is greatest, is 

 the same (as to its position of equilibrium) as if its centre of 

 gravity had been transferred about the ^j^th of an inch to- 

 wards the north end ; so that the moment of the force, 

 exerted by the vertical component of the earth's magne- 

 tism, can never exceed this small quantity multiplied by tlie 

 weight of the bar. Now, in order to render the results of 

 this instrument comparable to those of the horizontal-force 

 magnetometer, it should enable us to measure changes of 

 the vertical force, amounting to the ^^3___(|th part of the 

 whole; i. e. we have to measure effects, such as would be 

 produced by shifting the centre of gravity through the one- 

 millionth of an inch. It will be easily understood, from this 

 statement, how great must be the effect of a minute distur- 

 bance of the relative parts of ihe instrument, or of inequalities 

 in the bearing points of the axle ; and experience has ac- 

 cordingly shown that it is altogether unavailable for the de- 

 termination of changes of long period. 



The same difficulties, and from the same source, have 

 been found to attach to tiie usual method of observing the 



