CONTEMPORARY ADVANCES IN PHYSICS 343 



The desired photons being successfully fired into the gas, the next 

 problem is that of distinguishing the ionization they may cause in it 

 from whatever other liberation of charge they may effect in striking 

 walls, electrodes, or any of the other furniture within the tube. Light 

 of sufficient frequency to ionize a gas will usually be able to produce an 

 outflow of electrons from almost any metal. One takes of course the 

 elementary precaution of designing one's tube in such a way, that the 

 beam of light traverses it from entrance-window to exit-window 

 without touching any electrode; but the disparity of the effects is 

 nevertheless so great, that a modicum of scattered light may evoke 

 more electrons from the metal than the primary beam in all its strength 

 detaches from atoms of the gas. Some experimenters use alternatively 

 two electrodes of very unequal size, expecting that a current due to 

 ionization of the gas will be the same in magnitude whichever they use 

 as cathode, while a current due to light falling on the electrodes will 

 be greater when the larger is the cathode. Some vary the density of 

 the gas, assuming that if the current is proportional to density it must 

 be due to the effect which they seek; but it would also vary as the 

 density, if instead it consisted of electrons expelled from the electrodes 

 by light proceeding from excited atoms of the gas. Some finally have 

 so designed their apparatus that they perceive positive ions only; this 

 seems to be the safest way. 



Like Hughes,* I will divide the data according to the character of the 

 gases to which they refer: the common or "permanent" (chiefly dia- 

 tomic) gases first, then mercury, finally the alkali metals. 



The permanent gases can be disposed of in short order, for our 

 knowledge in this field is scanty, though surprising. Measurements of 

 ionization by electron-impacts, and what little has yet been deduced 

 from spectra, agree in indicating that for all of them (with the slight 

 exception of nitric oxide NO) the ionizing-potential is greater than 10 

 volts. Translating this figure into wavelengths of light, we infer that 

 only photons of smaller wavelength than 125w7/x can ionize such a 

 molecule in a single impact. Therefore light which is able to traverse 

 any window of solid substance should be unable to ionize any perma- 

 nent gas (except NO) ; in other words, any such gas enclosed in a tube 

 should be immune to ionization by any radiation entering from outside. 

 Yet there is unimpeachable evidence that air, and oxygen and nitrogen 

 separately, and possibly hydrogen and iodine, are ionized by light 

 which has penetrated fluorite. The threshold for these gases must 



* A. L. Hughes, Ionization of Gases and Vapors by Light (Washington University 

 Studies, 1929). 1 have benefited much by this article, and also by that of F. L. 

 iVIohler, Recombination and Photoio7iization (Reviews of Modern Physics 1, 216-227 

 (1929)). 



