Discharge of Electricity, 393 



The curve given by the mean equation is drawn on Plate XL 

 diagram 1. 



Here a and b are the semiaxes of the hyperbola represented 

 by the equation. The value of a is independent of the abso- 

 lute value of the entries; and as its different values agree well 

 with one another, we have here doubtless an important phy- 

 sical constant. I have also found its manner of dependence 

 on certain conditions. 



When hydrogen was substituted for air, and sparks taken 

 through it at the atmospheric pressure, I obtained a series of 

 readings which yield a result similar to that for air, namely 



V=43-19v/{s 2 + -1369s}, 



giving « = *0684 centim. and 5 = 2*954 C.Gr.S. units. Thus a 

 for hydrogen is less than a for air, the ratio being "66, which 



is very nearly the ratio of the - of the two curves (*65). 



When sparks were taken through air at the reduced pres- 

 sure of 180 millims., the curve obtained was 



V= 18-29^ + -5253s, 



giving a = *2616 centim. and b = 4*785 C.G.S. units. The 

 value of a was always found to be greater the smaller the 

 pressure. The distance at which the spark was first observed 

 to pass between the edges was 1*65 centim., compared with 

 1 centim. for the ordinary pressure. 



We found that it was possible to observe not only the 

 maximum deflection just before discharge, but also the deflec- 

 tion corresponding to a continued discharge produced by turn- 

 ing the plates rapidly. The deflection was always less than 

 that for the corresponding single discharge ; and the zero was 

 more displaced on the negative side of the original zero. The 

 wire-image can be made to remain very steady, with only a 

 slight oscillation. A discharge of this kind passed through 

 air at the atmospheric pressure gave a curve 



V = 45*58vV + -2046 s, 



which is similar to that for the single discharge; for the value 

 of a is the same, the only difference being in the diminished 



value of -. Part of this diminution may be due to the varia- 

 tion of the potential of the charged conductor; but I attribute 

 a considerable portion to a facility produced by the passage of 

 previous sparks. 



Having found the function for the difference of potential V, 

 we can deduce that for the electrostatic force R, and that for 



