Discharge from an Electrified Point, 



593 



The author found that these irregularities were due, at any 

 rate in part, to the presence of ozone in the discharge-tube, 

 because (1) they increased with the length of the experiment 

 and diminished again with a long wait, and (2) they were 

 reduced by placing a tray of powdered manganese dioxide 

 in the discharge vessel, to decompose the ozone which was 

 formed during discharge. Effects due to ozone could not be 

 completely eliminated in this way, because it is generated in 

 the path of the discharge itself, but the oxide no doubt 

 decreased the amount of it which was present at any given 

 moment. 



The effect on z of altering the size of the point and the 

 current was then determined. Three points were used, all of 

 platinum : two, A and B, had hemispherical ends and were 

 of diameter 0*078 cm. and 0*0043 cm. respectively. The 

 third, C, was a piece of the finest platinum wire cut obliquely 

 with scissors. Now the field at A was nearly four times as 

 great as that at B*; that at C probably many times greater 

 still. The values of z obtained were, however, nearly the 

 same for all. 



Thus the values of z for A were 0'43 cm. in positive dis- 

 charge, and 0*4:0 cm. in negative discharge. The corre- 

 sponding values for B were 0*44 and 0*41, and those for C 

 were 0*39 and 0*36. 



Those values were also independent of the current, as will 

 be seen from the following table, if experimental discrepancies 

 are allowed for. 



c 



(microamperes 

 per sq. cm.). 



*o + . 



Zq— . 



0-05 



•45 



•40 



0-14 



•47 



•38 



0*14 



•38 



•28 



0-31 



•43 



•49 



031 



•42 



•40 



0-69 



•45 



•35 



117 



•43 



•43 



•44 



•40 



* Obtained from the expression fr° '^=^ constant, Phil. Mas:, xx. p. 270 

 (1910). 



