s<mi-: C(K\rr.Mi'(V<.ih'y .-inr.ixcrs tx rinsus ri \m 



iur iIr- furious luMt ik-M-Iopi'd ,il tlu- ciilindf wluti tin- inns lin,ill>- 

 strike it. Just ln-lori' tlu- ions i»rri\o ,it ilu- i.iiIkkIi' thr\ must \k- 

 t'lulowi'd with a kiiu-lic oniTk;\' which is \i'i\ uinisu.il (lits.i\ ihi- iiMsl ) 

 in tlie miililio of iht- ilisrli.irm-; and it is in fa( i ol)si'rvfd thai just in 

 front of tlie cathodt- there is a sharj) and sudden jiotenti.d-f.dl, cor- 

 res|>ondin); to a strong field extending; I)ut a little \va\- outward from 

 the electrode and then dsing down into the weak lield pre\ailinK 

 throusli the rest of the arc. This strong lield picks up the ions which 

 have meanderetl to its outward edge from tiie hody of the discharge 

 and hurls them against the cathode — not very forcihh , for the energy 

 they recei\e from that potential-fall is not a great amount by ordinary 

 standards, and most of the ions probably lose some of it in collisions 

 on the way; but with much more energy than they would be likeK' 

 to possess anywhere else in the arc. 



This fX)tential-fall immediately in front of the negative electrode, 

 the cathode-fall of the arc, is measured by thrusting a probe or sound- 

 ing-wire into the discharge as close as possible to the cathode (gener- 

 ally about a millimetre away), and determining the P.D. between 

 it and the cathode. The probe is regarded with some distrust, as it 

 raises in an acute form the old ciuesiion as to how far the phenomena 

 we observe in nature are distorted by the fact that wc are observing 

 them; the wire may alter the potential of the point where it is placed, 

 or it may assume a potential entirely different from that of the en- 

 vironing gas; but the general tendency nowadays, I believe, is to 

 accept its potential as a moderately reliable index of the potential 

 which would exist at the point where it stands if it were not there.'" 

 The cathode-fall, as so measured, depends unfortunately on quite 

 a number of things; the material of the cathode, the gas, the current. 

 The gas is always mi.xed with a vapour of the electrode-material, 

 particularly in the vicinity of the electrode; the only way to have a 

 single pure gas is to enclose the whole system in a tube, evacuate the 

 tube to the highest possible degree, and then heat it until the vapor- 

 tension of the metal of which the cathode is made rises high enough 

 for the vapor to sustain the arc. This is practicable with the more 

 fusible metals; and with mercury, the arc generates heat enough to 

 maintain the vapor-tension sufficiently high. In pure mercury- 



" On this matter the e.xperiments of Langniuir and Schottky, mentioned further 

 along in this article, promise new knowledge. The probe automatically assumes 

 such a potential that the net current-flow into it is nil; for example, if it is immersed 

 in an ionized gas in which electrons and ionized atoms are roaming about, its eventual 

 potential is such that equal numbers of particles of the two kinds strike and are 

 absorlwd in it per unit time. If the electrons are much more numerous or have a 

 much higher average energy, or both, this potential may be several volts more 

 negative than the potential at the same point before the probe was put in. The 

 same may be said about the wall of the tube. 



