300 BELL SYSTEM TECHNICAL JOURNAL 



ha\c a single energj' conforming to equation (1)? One could not 

 affirm this a priori, for a solid metal is not a collection of free atoms 

 close together as a gas is an assemblage of free atoms far apart, but 

 rather a structure of atoms which interfere with one another and are 

 distorted, and there are many electrons in a solid of which the bonds 

 and the constraints are very different from those by which the elec- 

 trons of free atoms are controlled and vice versa. When a plate of 

 sodium or a pool of mercury is exposed to a rain of electrons, not 

 exceeding say 10 equivalent volts in energ>', nothing apparent hap- 

 pens." When the vapor of either metal is similarly exposed, the atoms 

 respond in a manner from which they are inhibited, when they are 

 bound together in the tight latticework of a solid or the promiscuous 

 crowding of a liquid; and light is emitted. 



The phenomena are clearest when the bombarded \apor is that of 

 a volatile metal, such as mercury, sodium, or magnesium. The atoms 

 in such vapors are not usually bound together two by two or in greater 

 clusters, as they are in such gases as oxygen or hydrogen, of which 

 the response to electron-impacts or to radiation is not quite under- 

 stood to this day; and the first radiations which they emit arc not 

 in the almost inaccessible far ultra-violet, like those of the monatomic 

 noble gases, but in the near ultra-violet or even in the visible spectrum. 

 Dealing with such a vapor, I will say mercury for definiteness, one 

 observes that so long as the energy' of the bombarding electrons 

 remains below a certain value, no perceptible light is emitted; but 

 beyond, there is a certain range of energies, such that electrons pos- 

 sessing them are able to arouse one single frequency of radiation 

 from the atoms. Ordinarily, as when a \apor is kept continuously 

 excited by a self-sustaining electric discharge throughout it, the 

 atoms emit a great multitude of different frequencies of radiation, 

 forming a rich and complicated spectrum of many lines. But if the 

 energy of the bombarding electrons is carefully adjusted to .some value 

 within the specified range, only one line of this spectrum makes its 

 appearance; under the best of circumstances this single line may be 

 exceedingly bright, so that the absence of its companions— some of 

 which, in an ordinary arc-spectrum, are not much inferior to it in 

 brightness — is decidedh' striking. The one line which constitutes 

 this single-line spectrum is the first line of the principal series in the 

 complete arc-spectrum of the element; its wave length is (to take a 

 few examples) 2o3GA for mercury, .'jSOO for sodium (for which it is a 

 doublet), 4.571 for magnesium. 



' According to a vcr>' recent paper by C. H. Thomas, radiations from iron excited 

 b>- electrons with as low an energy as some two or three equivalent volts have been 

 detected. 



