162 Dr. W. F. G. Swann on the Magnetic Field produced 



consisted in overcoming the various causes of fluctuation due 

 to direct electrostatic effects, variations of the earth's field, &c, 

 and once these had been successfully overcome, it was felt 

 that multiplication of the observations could not add anything 

 to the reliability of the results. 



Method of measuring the quantity V referred to on page 158. 



The method adopted was to measure the potential at 

 various points along the line joining the inner point of inter- 

 section of the coils with the paraffin wax, then by plotting 

 the potential against the distance from one end of this line, 

 the quantity V representing the mean value of the potential 

 could be obtained. 



In order to measure the potential at any point along the 

 above line a special method was adopted, since the intro- 

 duction of anything of the nature of a wire connected to an 

 electrometer of any appreciable capacity would result in such 

 an induced charge on the wire as would completely alter the 

 potential to be measured. 



A small proof plane (about 3 mms. in diameter) was 

 placed at the point at which the potential was required, and 

 was then touched by a thin earth-connected wire. The proof 

 plane received an induced charge, and its potential became 

 zero. Now assuming that the induced charge did not appre- 

 ciably alter the distribution of electricity on the charged 

 plate, and on the coil in its vicinity, which assumption was 

 quite justified to the degree of accuracy to which we were 

 working, we could look upon the potential of the proof plane, 

 which was zero, as being made up of two equal and opposite 

 parts: (1) the positive potential which would have existed at 

 the point at which we desired the potential, if the plane had 

 been absent, (2) the negative potential which the charge on 

 the proof plane contributed. If the earth-connected wire 

 were then removed the potential of the proof plane would 

 still be approximately zero, except in so far as the small 

 amount of charge on the wire itself contributed to the 

 potential of the plane *. If the proof plane were now re- 

 moved and held away from the apparatus, it would have a 

 potential as much below zero as the potential which we 

 desired to measure was above zero. It only remained to 

 measure the potential of the proof plane. 



For this purpose an electrometer of small capacity was 

 constructed. It consisted of a silvered glass fibre, with a 



* The error due to this cause became automatically eliminated in 

 practice (see note on page 163). 



