On Magnetism, Cohesion, Adhesion mul Viscosity. 44-7 



of no consequence, as anything lower than the azimuth line 

 that passes through 5°, I have found to be as low as can be 

 ever required for any useful measurement; and it will be seen 

 that by the dotted line I have drawn passing through it, and 

 the dotted lines that are intended to represent the correction 

 of the curve, that azimuth remains unaffected. Measured by 

 a rule, the proportion of the azimuth line passing through 90° 

 is as 1 to 2 with that passing through 5'~'. Now by the inverse 

 ratio of the distance, it has only one-fourth of the strength at 

 5° that it has at 90° ; and practicall}', therefore, is the same in 

 strength at 5° as it is at ITS'". Hence 90° may be considered 

 as a middle point, from which the degrees diminish by equal 

 decrements to 5° and 175°; and it will be seen by a reference 

 to the curve, that the azimuth lines cutting 45° and 135° are 

 essentially the same in length. The ascertainment of the true 

 strength of the degrees from 90° to 180° was a necessary de- 

 sideratum, which has led to the unexpected discovery that the 

 position of the magnetic needle north and south is the effect 

 of equilibrium arising from equal repulsions from east and 

 west ; and that so far from there being any attraction at 

 N. and S., they are just the two points where the needle is 

 in the greatest state of freedom ; and that in the case of a nee- 

 dle of low magnetic power, the mere resistance caused by 

 friction on its pivot, will be sufficient to prevent its taking up 

 its true position in the line of the magnetic meridian. These 

 facts prove that the movement of the needle from 90° to zero 

 depends ujion repulsion, commencing at 90°, and diminishing 

 gradually till it stands in equilibrium at zero ; and that the 

 return of the needle to 90'', when the north end is brought to 

 ISO'', is the result of an attraction that increases up to 90°, 

 when the repulsion commences. 



The mathematical expression of these results is that — 



The value of every degree of the compass is inversely as the 

 square of the length of the ordinate or co-ordinate -passing 

 through it, the abscissa bei7ig considoed as zero. 



The general principle of the values of the various decrees 

 of the compass obtained by the preceding system of azimuths, 

 was further verified by the following method : — 



A bar- magnet 12^ inches long, was properly suspended by 

 silk, free from torsion, and the N. pole being brought round 

 to the south, was allowed to return back by the east. The 

 following is the mean of eight observations, made with great 

 care, and which only slightly differed from one another. The 

 bar required 11,} seconds to complete its return from 180° to 

 135 . When made to start again from 135° to 90° it occu- 

 pied six seconds ; and when allowed to depart again from 90° 



