910 



SCIENCE 



[N. S. Vol. XLIV. No. 1148 



tion. The bands of some substances, such 

 as nitrogen, are not found in absorption 

 under any conditions, and the conditions 

 of their occurrence indicate that the emis- 

 sion bands are due to the recombination of 

 a detached electron with a positive mole- 

 cule. The negative pole bands appear 

 under the same conditions as spark lines, 

 and it seems not improbable that they are 

 due to the neutralization of a doubly 

 charged molecule. 



The spectral differences attending dif- 

 ferent stages of ionization are well illus- 

 trated by some recent experiments. 

 Franck and Hertz found that mercury 

 vapor is ionized by a field of 4.9 volts, and 

 then emits the one ultra-violet line 2537. 

 The Einstein relation Ve = hv is fulfilled. 

 McLennon and Henderson verified this con- 

 clusion, and also found that with a field 

 of about 12 volts a second stage of ioniza- 

 tion occurs, attended by the emission of 

 the many-lined spectrum attributed by 

 Stark to multiple charges. McLennon 

 finds that zinc, cadmium and magnesium 

 also give single line spectra which prob- 

 ably conform with Einstein's equation, 

 which we should not expect to apply in a 

 simple form to the many-line spectrum. 



It appears from such experiments that 

 there is a threshold value of kinetic 

 energy which must be imparted to an emis- 

 sion center before it can radiate, which 

 represents the work of ionization and is 

 equal to a light quantum. Franck holds 

 that this energy may be devoted either to 

 ionization or to emission, but that both can 

 not simultaneously occur. Stark believes 

 that the two are coincident, the emission 

 accompanying the rearrangement of elec- 

 trons in the atom after one has been 

 ejected. This suggests an explanation of 

 quantum emission involving no departure 

 from accepted electromagnetic theory. 



The spectra of hydrogen and of helium 

 are of particular interest because their 



atoms are of the simplest type and because 

 it is possible that they are the basic units 

 of which all elements are composed. The 

 Pickering series in stellar spectra was at- 

 tributed to hydrogen because of its nu- 

 merical relationships with the Balmer 

 series. The study of series relations led 

 Rydberg to predict the 'occurrence of a 

 principal series for hydrogen beginning at 

 wave-length 4686, and this line was subse- 

 quently found in nebular and stellar spec- 

 tra. After many attempts to reproduce 

 these spectra in the laboratory, Fowler suc- 

 ceeded in 1812, by passing a powerful dis- 

 ruptive discharge through a mixture of 

 hydrogen and helium. Produced only 

 under such conditions, these must be 

 classed as spark lines ; and if Stark 's views 

 are correct and if they are really due to 

 hydrogen, that element must have more 

 than one detachable electron. 



In applying his theory to the helium 

 spectrum, and assuming one electron re- 

 turning to a helium atom from which two 

 electrons have been detached, Bohr ob- 

 tained a formula which gives lines corre- 

 sponding in position to those of the Pick- 

 ering and Rydberg series, and also another 

 series almost coincident with the Balmer 

 hydrogen series. This remarkable conclu- 

 sion was strengthened by Stark's discov- 

 ery of 4686 in a helium tube which gave no 

 lines of the ordinary hydrogen spectrum. 

 He concluded from the canal-ray displace- 

 ments that the emission centers were 

 doubly charged. Evans also found the 

 first members of all the series assigned to 

 helium by Bohr, including that correspond- 

 ing to the Balmer series, in a tube contain- 

 ing no hydrogen. The experimental evi- 

 dence thus favors Bohr's theory, but we 

 must remember the remarkable way in 

 which the presence of one element may in- 

 tensify or suppress the spectrum of another. 

 For example, Lyman found that the ultra- 

 violet series attributed without question to 



