1904.] 



between Electrically Charged Surfaces. 



339 



I have investigated this subject by means of the electric micrometer,*" 

 which is specially well adapted for the purpose, as it measures by the 

 principle of electric touch. Briefly expressed, it has a fine micrometer 

 screw S (fig. 1) with a graduated wheel G ; and a system of six levers is 

 arranged to minify the movement of the screw point, thus 1 /x movement 

 of the screw-head becomes 1 /x/x or less of the point X. A fixed surface 

 Y is arranged near the adjustable one X, and when X and Y touch a 

 circuit is completed as shown through the telephone T. To find the 

 position of contact the observer works the screw S, watches the wheel 

 G, and listens at T. The spark gap is across XY. A battery B 

 capable of giving 200 volts has a potential divider Z h which enables us. 

 to apply to XY any required voltage, which is read on the voltmeter 

 W. A key K enables us to use either the large voltage from B on a 

 small voltage volt) from A, also provided with a potential divider 

 Z 2 . To make an observation apply the small voltage from A, ascertain 

 the position of contact by G and separate the contacts XY. Now apply 

 the large voltage from B and slowly bring the contacts together until 

 discharge occurs, as observed in the shunted telephone. G is again 

 read. The results follow (p. 341). 



The table shows Earhart's results with mine. There are some- 

 points of interest : — 



1. The last two columns show potential gradients. There does not 

 appear to be any change in gradient nearer the origin than 2 /x, hence 

 we infer that there is no inner film or change in dielectric strength. 



2. In working the electric micrometer I commonly use a potential 

 difference across the contacts of -^-th volt. ; it might be supposed that 

 this voltage would produce irregular discharge and make the readings 

 of the instrument uncertain. The table shows that the sparking 

 distance would be about \ fxfx. So that even if there were a large 

 percentage error in this distance the errors in measurement would be 

 insignificant. 



The curves are shown in fig. 2. The curves on the left side indicate 

 good agreement between the two observers, where they overlap. 



We should expect more regular results when there is a pure liquid 

 between the surfaces than when there is an air gap ; for dust particles 

 which would interpose in the air could not break into the liquid.. 

 Earhart's results show great difference in regularity on the two sides 

 of the " knee." The curves on the right are on an open scale to show 

 my results better. 



The potential gradient for Earhart's results is 200 — for mine 150,, 

 the curves crossing one another for voltage 60. 



* For an account of this instrument, see Shaw, ' Phil. Mag./ December, 1900 j. 

 March, 1901; 'Electrician,' February, 1901. 



