INTRODUCTION. 61 



the axis of the needle turn on friction-wheels ; a 

 contrivance which, however, has been rejected, as 

 useless, by the latest observers, Humboldt, Biot, 

 and Tobias Mayer, and exchanged for very fine 

 axes running in agate holes. In order to remove 

 the influence of an imperfect balancing of the 

 needles, Cavendish recommended to furnish them 

 with moveable weights, by the due distance of 

 which from the centre, the needle, merely by the 

 effect of gravity, would repose pretty nearly in the 

 inclination which would be produced by the mag- 

 netic power in that latitude, that magnetism might 

 the more easily conquer the other obstacles. But 

 not to mention that, by such a contrivance, we are 

 exposed to the suspicion of having ourselves pro- 

 duced the phenomenon which we mean to observe, 

 the balancing of those weights at every consider- 

 able change of latitude is no business for the sea- 

 man J and, in particular, the careless manner in 

 which those weights are fastened, and exposed to 

 every accidental disturbance, is objectionable. 

 Under these circumstances, there is no general 

 and certain method of finding the true dip of the 

 magnetic needle, but to turn the needle round, 

 and then, by the help of a strong magnet, to re- 

 verse its poles, and so to observe the dip in four 

 different positions. The first two observations give 

 the place of the point of gravity of the needle in 

 the vertical direction (in the breadth of the 

 needle) ; the two last, its distance from the centre 



