Magnetism, Cohesion, Adhesion and Viscosity, 445 



tion at eight inches distance, owing to the minute quantities 

 concerned, I could obtain a deviation at sixty inches, after 

 which all effect ceased. As the distance was exactly four 

 inches from the centre when the north end of the needle was 

 (h-awn to 180° by the bar-magnet, and as there must be some 

 definite ratio between the two distances thus obtained, I as- 

 sume that the true length of the greater azimuth is sixty-four 

 inches, and that the four inches not discovered in the measure- 

 ment resulted from the friction of the needle on its pivot, 

 though it was an extremely delicate instrument that had been 

 made by M. Pixii. Instead therefore of the greater distance 

 being merely the duplicate length of the shorter one, it is re- 

 lated to it as its cube. 



The reason of this disproportion of the needle when in its 

 natural position, from that which is discovered when it is 

 forced round to the opposite pole, may, I think, be found, 

 when it is remembered that, as the earth is either per seovhy 

 induction a magnet, the latitude in which the experiment has 

 been made being at such a distance from the magnetic equa- 

 tor, there can be no longer an equality in the effects. It would 

 be therefore a question, I think, not devoid of interest, to as- 

 certain whether such a disproportion exists at the magnetic 

 equator. 



I will now describe the means by which a long line was 

 obtained, upon the exactness of which the accuracy of the 

 observation necessarily depended. A white thread of fine 

 cotton, weighted at each end with lead, was laid over a long 

 and steady table, and the weights being allowed to hang down, 

 a perfectly right line was instantly obtained, which had only 

 to be adjusted so as to agree with the magnetic meridian. 



The length of the azimuths obtained in the lines N., E. and 

 S. being respectively 64, 2, 4, it follows that by the law of the 

 inverse square of the distance they are related to one another 

 in the proportion of J •1024 and 256, and that consequently 

 every degree rises gradually from N. or 0°, till at E. or 90° its 

 strength is 1024, after which there is a rapid decrease up to 

 180°; and it may therefore be practically assumed, to save 

 circumlocution, that at 175° its proportional strength is 256°, 

 or only one-fourtii tiie strength it has at 90°, that being the 

 true proportion as it readies 8. or 180^ where the needle is 

 necessarily in a state of equilibrium by its balanced attraction 

 to the west; but all this will be much more evident by a re- 

 ference to tile curve itselfj whicii is preserved in the form that 

 was laid belbre the Royal Society, as it would be inconvenient 

 to represent on paper the due proportion from the centre to 

 N. In fact, for ail practical purposes, the new correction is 



