Magnetism on a permanent Electric Current. 309 



nest films used by measuring the electrical resistance ; but this 

 method, as will be seen, is exceedingly faulty. The thicker 

 strips have been weighed before being placed on the glass; but 

 even this method fails to determine the effective thickness accu- 

 rately. Even if the specific gravity were the same for all the 

 strips (and it probably is not), the value thus obtained for* the 

 thickness would give only the average thickness ; and this is 

 by no means the effective thickness. It will be remembered 

 that the connexions leading to the Thomson galvanometer are 

 placed opposite to each other, with the width of the metal strip 

 between them. The effective thickness is the average thick- 

 ness along the line joining these two side connexions. Gold 

 foil is obtained in sheets 10 or 12 centims. square. It will be 

 seen further on that in one case two strips cut from similar 

 positions in the same sheet differed in average thickness about 

 7 per cent. This being the case, it seemed quite possible that 

 the effective thickness of any strip, as defined above, may 

 differ many per cent, from the mean thickness indicated, by 

 the weight. 



All these sources of error being considered, the discrepan- 

 cies which will be observed in the results to be given will not 

 be surprising. 



A single complete series of observations consisted of the fol- 

 lowing parts : — 



1st. A determination of the extent to which the indicator of 

 the Thomson galvanometer was affected by] the direct influ- 

 ence of the magnet and the magnetizing current. — All that it 

 was necessary to ascertain in this case was the change in 

 position of the galvanometer indicator caused by reversing 

 the current through the magnet. This usually amounted to 

 1 or 2 millims.; and subsequent readings of the Thomson 

 galvanometer were, when it was necessary, corrected accord- 

 ingly. 



2nd. A determination of the strength of the magnetic field. 

 — This was done by withdrawing suddenly from the field a 

 small coil consisting of a few turns of wire and observing the 

 effect of this action on a delicate galvanometer placed in cir- 

 cuit with the coil *. The galvanometer was used with a mirror 

 and scale ; and the readings actually obtained were reduced by 

 the formula 



2~4A 2W P 



where n is the actual reading and r the distance from the 

 mirror to the scale. The constant of the galvanometer not 



* Rowland, " On a Magnetic Proof Piano," Silliinan's Journal, vol. x; 

 p. 14 (1875). 



sin 



