344 BELL SYSTEM TECHNICAL JOURNAL 



therefore lie on the long- wave side of 125m/x. For air it is presumed to 

 lie between this and 145m/x, since a plate of quartz holds back the ioniz- 

 ing rays. The discrepancy between these and the expected thresholds 

 may not seem large, but it is important. Naturally one has recourse 

 to the idea of a two-stage ionization, occurring when two quanta in 

 succession are absorbed by the same molecule— an idea which, we 

 shall see, is frequently invoked in other cases. If this is valid, the 

 ionization should increase as the square of the intensity of the light. 

 There seem to be no data bearing on this point. To quote from 

 Hughes, "further investigations in this field are badly needed."^ 



With mercury the situation is much more definite, but for those who 

 like to have simple theories verified completely it is no more satisfac- 

 tory. 



The spectrum of the mercury atom is well mapped and well inter- 

 preted, and the ionizing-potential for electron-impacts has been deter- 

 mined over and over again. From both of these it follows that ioniza- 

 tion by single photons should be possible only at wavelengths smaller 

 than 1188A. However it is certain that the light of the famous res- 

 onance-line of mercury, 253 7A, is able to ionize the vapor of the ele- 

 ment whence it proceeds.- 



This seems the natural equivalent of the well-known fact that when 

 mercury atoms are bombarded by a sufficiently dense electron-stream, 

 ionization begins at the resonance-potential. The quanta of the 

 wavelength 2537A have 4.9 equivalent volts of energy. Such a 

 photon strikes an atom, and excites it transferring if from the normal 

 into a certain excited state, denoted by the symbol 2^Pi; a second 

 comes along and likewise is absorbed, bringing the energy of excitation 

 of the atom up to twice 4.9 equivalent volts; this amount falls short of 

 the ionizing potential by only 0.6, and a third photon more than sup- 

 plies what is required. So runs the simple interpretation; but we shall 

 see that only the first of these steps is confirmed by further experiment, 

 and that the rest of the process must happen in some other way, 

 though the way is far from clear. 



1 For references to the literature I refer to Hughes {I.e.) Among the latest papers 

 are those of A. L. Hughes {Proc. Camb. Phil. Soc, 15, pp. 483-491 (1910)); F. Palmer 

 (Phys. Rev. 32, pp. 1-22 (1911)); E. B. Ludlam {Phil. Mag. (6) 23, pp. 757-772 

 (1912)); W. West, E. B. Ludlam {Proc. Roy. Soc. Edinb., 45, pp. 34-41 (1925)). 

 Some of the early work on air was confused by what appears to have been a photoelec- 

 tric effect of particles of colloid size ("nuclei" or Kerne) produced by the action of 

 ultraviolet light on impurities in the air — one of the once-popular and now forgotten 

 problems of physics. 



2 Literature: G. F. Rouse and G. W. Giddings, Proc. Nat. Acad. Sci., 11, pp. 

 514-517 (1925); 12, pp. 447-448 (1926). F. G. Houtermans, ZS. f. Phys., 32, pp. 

 619-635 (1925). Twenty years ago W. Steubing {Phys. ZS., 10, pp. 787-793 (1909)) 

 observed that light coming from a mercury arc and passing through quartz was able 

 to produce a current in a tube containing mercury vapor; but his result has been 

 impugned. 



