THE PEESSURE UPON THE POLES OF THE ELECTRIC ARC. 
135 
Ihe method of starting an electric arc by forcing a spark between the separated 
poles possibly depends as much upon the photoelectric action induced as upon the 
ionization within the spark gap. 
Polar Lines in Arc Spectra. 
The original experiment upon the pressure upon the poles on which the above 
research is based was carried out in 1912, and had been undertaken in the expectation 
of finding a recoil effect. The writer had previously described a series of spectrum 
lines which made their appearance near the poles of an iron arc, to which the name 
“ polar lines had been given,* and in the discussion of their origin something in the 
nature of an explosion upon the surface of the pole was suggested to account for the 
potential drop with which they appeared to be associated. 
It seems now possible to go further than I did in the original paper and state that 
the explosion results in the liberation of an electron with high speed, and it is further 
suggested that the polar line is due to the particular type of vibration which is set 
up at the instant when an electron is expelled from the atom. The other, or median, 
lines being due to the secondary action when the electron with its ionizing velocity 
impinges upon another atom. The feeble intensity of the polar line contrasted with 
the greater intensity of the median line is in accord with this view. 
A feature of the occurrence of polar lines is their predilection for regions of the 
spectrum of short wave-length. In the case of the iron arc they became increasingly 
numerous as more refrangible parts of the spectrum were reached. It is quite 
possible that their preference for this range is due to their photoelectric origin, since 
such action is limited to regions of high frequency; in the rare case (only as far as I 
know in the iron arc) in which a small grou]^ is found in another part of the spectrum 
the effect may be a resonance one. 
Rossif has observed polar lines in the spectrum of the copper arc, and it appears 
that 2714 A.U. is the wave-length of the least refrangible one. Richardson and 
Compton^ give 3000 A.U. or 3090 A.U. as the longest wave-length capable of 
producing photoelectric emission from copper. There is thus further evidence of the 
polar lines being due to photoelectric action. In the case of iron the polar lines were 
found to be nearly equally strong at the two poles, but with the copper arc Rossi 
found that unless the current was strong they were confined to the cathode; this 
suggests that the attraction of the anode for the electron was sufficient to prevent its 
expulsion. 
The reason why an explanation based upon photoelectric rather than thermionic 
action is offered is that in the experiments already quoted it was found that the 
* Duffield, ‘ Astrophysical Journal,’XXVIL, 264, 1908. 
t Rossi, ‘ Astrophysical Journal,’ XXXV., 279, 1912. 
1 Richardson and Compton, ‘ Phil. Mag.,’ XXIV., 575, 1912. 
