Galvanometric Method of Measuring H. 81 



The lower portion of the figure represents the potentiometer ; 

 it would, of course, be preferable if this could be replaced by 

 a battery of sufficiently constant E.M.F. capable of giving a 

 milliampere without polarizing. Not knowing of any such 

 battery I was compelled to use accumulators, which will give 

 quite large currents, and vary less than 10 ^ 00 part in half 

 an hour, and measure the E.M.F. at each observation. 



Gr was generally a tripod galvanometer of about three 

 ohms' resistance ; with the voltage used, R had to be made 

 about 4000 to give a deflexion of 45°, so that the current 

 used never much exceeded a milliampere ; this was quite 

 insufficient to change the temperature of the coils appreciably. 

 The resistance of G and the leads to it being a very small 

 fraction of R, their temperature-variation was of no account, 

 and only that of R in resistance and C in E.M.F. had to be 

 considered. 



The usual procedure was to set the galvanometer Gr so as 

 to bring the spot of light to zero ; then turn it 45° one way, 

 and apply a current, adjusted by the box R, sufficient to bring 

 the light back to the zero ; then take the potentiometer 

 reading and temperature ; break current ; set the galvano- 

 meter at 45° in the other direction from the meridian ; apply 

 the necessary current, the key K being reversed : while this 

 flowed, read the potentiometer again, then break the current, 

 and turn Gr into its original azimuth, noting that the light 

 returned to the exact position. The whole operation takes 

 scarcely five minutes. Sometimes a longer series of observa- 

 tions was taken, with deflexions of 15°, 30°, 45°, 60 c , 75°, on 

 each side of the meridian, and 90° on one side. The needle 

 is never stable at 90° on both sides at once, as stability 

 there depends on the direction of the torsion. Pairs of 

 observations at 90° apart were taken, and equation (3) applied 

 to calculate H and r(f>. The following is a specimen of the 

 results obtained : — 



H 

 (in arbitrary units) . 



Deflexions. 





o 



+ 75 



and 



o 



-15 



+ 60 



V 



-30 



+ 45 



j? 



-45 



+ 30 



}? 



-60 



+ 15 



•» 



-75 



mean 8155 



8178 

 8154 " 

 8155 

 8156 

 8156 m 



The first result is untrustworthy, because for S=— 15° R waa 

 beyond the range of the resistance-box and had to be ton ml 

 by extrapolation ; the others agree well. 



I made no attempt to reduce the results to absolute measure, 

 Phil. Mag. S. 5. Vol. 33. No. 200. Jan. 1892, G 



