Force of the Electric Discharge through Gases. 437 



the string on which it acts, Ave have the following formula 

 connecting the tension T along the discharge with r, the 

 radius of curvature of the deflected path, 



T=F.r, 



where F= the force deflecting the path of the discharge at 

 any point 



=c/, 



where C is the current in the discharge, and /is the magnetic 

 force at right angles to the discharge. 



If C, r, and /are expressed in C.G.S. units, T in the above 

 formula will be given in dynes. 



On calculating T, however, for different values of.(^ft-n# 

 obtained by experiment, the values were found to .differ widely 

 for the same gas. $r 



77ie JSature of the JDejieAon hy Magnetic Force. 



From the behaviour of the discharge when the magnetic 

 field was suddenly made, or being made was quickly reversed 

 in direction, it seemed more probable that the deflexion of the 

 discharge by magnetic force was due to the same cause as 

 that produced by the convection-currents in the gas, and that 

 in either case the path of the discharge was determined by the 

 line of least resistance through the gas between the electrodes. 

 The discharge could hardly be deflected in the form shown in 

 figs. 6, 7, or 18 if it were held in position by a tension along 

 the line of discharge. 



It is well known that the electromotive force required to 

 maintain a discharge in a rarefied gas is not nearly so great 

 as that required to start it, since the molecules of the gas 

 dissociated or ionized by the first discharge leave, as it were, 

 a thread of conducting material between the electrodes, along 

 which succeeding discharges pass much more easily than 

 through fresh gas. When this thread of ionized gas is dis- 

 placed by any means — e. g. magnetic force acting on the 

 charged and moving particles of the gas, or by currents in the 

 gas itself due to any cause and blowing across the electrodes — 

 the discharge follows it, since the increase of the resistance 

 due to the increase in length of the displaced path is more 

 than compensated for by the greater conductivity in this 

 direction. 



If, however, the time between successive discharges is so 

 great that the thread of ionized gas produced by one discharge 

 becomes dissipated, or is destroyed in any way, before the 

 next discharge takes place, then we should expect no deflexion 



