80 Mr. R. A. Lehfeldt on a 



of the meridian, representing these hy the suffixes 1 and 2, 

 we have 



Yx G + T^-Hsin(a i + ^)=0, 



-y 2 G + T0 + Hsin(S 2 -<9)=O, 

 whence 



H^CP^ + Y^ + StV + ^G^-Ys) . • (3) 



By taking S 1 = S 2 we shall get Yi nearly equal to Y2? and the 

 last term will become almost negligibly small ; the middle 

 term is always so. But it is better to use such a pair of 

 observations merely as a preliminary test, to indicate the 

 direction of the torsion, so that it may be removed. I found 

 that, using silk fibres, after turning the torsion-head, the 

 results obtained were at first very irregular ; but after 

 successive attempts had reduced the angle of torsion to a small 

 amount, and the fibre had been in use for some days, the 

 elastic after-effects disappeared. So that in a magnetic 

 observatory, where the instrument was always kept for the 

 same purpose, a silk fibre could be used, and once well freed 

 from torsion would probably work quite satisfactorily. The 

 only precautions necessary to note are that, when the needle 

 is deflected by a current, the galvanometer should, without 

 delay, be turned to bring the spot of light back to zero ; and 

 that the needle-chamber should be small, not only because 



Laa/wwwmAAM/v 



L-^ — till * 



H), 



it is convenient to work with well-damped vibrations, but in 

 order that no accidental shaking or current should turn the 

 needle round 360°. 



The apparatus was arranged as shown in the diagram : — 



B, battery of two storage cells. 



C, Clark cell. 



G, the principal galvanometer. 



H, auxiliary galvanometer for determining the E.M.F. 



K, four-part plug-key. 



R, adjustable resistance. 



S, potentiometer wire-bridge. 



