335 



laws both of the variation of the force with which a truly suspended 

 magnet is directed, and of the direction of that force itself. 



The most convenient of these is that termed the declination 

 magnet, which is suspended horizontally by a fine silken thread ; the 

 tendency of the needle to dip being obviated by placing the point 

 of suspension north of the centre of gravity. This instrument is 

 very convenient, especially in high latitudes, for exhibiting in a mag- 

 nified form that portion of the motion of the freely suspended dip- 

 ping-needle, which is at right angles to the vertical plane of the 

 needle. Two other instruments, one termed the bifilar magneto- 

 meter, from its suspension by two threads; the other named the balance 

 magnetometer, from its resemblance to the beam of a balance, en- 

 able us to observe the variations of the horizontal and vertical com- 

 ponents of the force with which the freely suspended dipping-needle 

 is directed ; whether these variations be due to a change in the total 

 value of the force, or simply to a change in its direction parallel to 

 the vertical plane. In high magnetic latitudes, the bifilar or hori- 

 zontal component magnetometer will be most affected by changes of 

 the direction of the force in the vertical plane, and the balance or 

 vertical component magnetometer will be most affected by variations 

 of the intensity of force : in low latitudes the reverse is the case. In 

 all three instruments the magnets are forced from their natural posi- 

 tion. By means of a well-known formula, however, we can com- 

 pute, from the observed variations of the two components,the variations 

 of the total force, andof its direction in the plane of the magnetic meri- 

 dian. Theoretically this operation is simple enough, but practically 

 there are great difficulties; these difficulties are due to the effect of tem- 

 perature upon the positions of the bifilar and balance magnets, which 

 require to be eliminated, and to sources of error that I have pointed 

 out in the Edinburgh Transactions in the determinations of the 

 change of value of either component of force, which corresponds to 

 a change of, say one minute in the angular positions of the magnets. 

 I conceive that I have, by the employment of new methods, reduced 

 the errors due to these causes to a very small amount ; and it is for 

 this reason that I claim for the results deduced from the Makerstoun 

 Observations, a consideration which they could not otherwise have 

 been entitled to. I refer to the part of the Transactions now in the 

 press, for the results relative to the separate magnetic elements, and 

 to thL total force; I confine myself at present to those touching the 



