906 



SCIENCE 



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



field. The inverse absorption effect has so 

 far not been observed. 



Long before the Stark effect was ob- 

 served Voigt showed that such results 

 might be expected from quasi-elastic 

 forces in the atom and the stresses pro- 

 duced by the field. Schwarzschild has at- 

 tempted to explain it by the ordinary laws 

 of electrodynamics, and Warburg, Gehrcke, 

 Garbasso and Bohr by Bohr's theory. 

 Each attempt was successful in some re- 

 spects, but each failed to account fully for 

 all the components, their displacements 

 and their state of polarization, and all the 

 theories assign the same number of com- 

 ponents to each line of a series, whereas 

 one of the most significant features is the 

 progressive difference in number of com- 

 ponents, displacements and relative inten- 

 sities in passing from one line to another. 

 Stark not only rejects them all, but is led 

 by his study of the phenomenon to finally 

 abandon the quantum and light-cell theo- 

 ries, because he considers that he has 

 proved that the greatest possible energy 

 which an electron can acquire in its orbit 

 falls far short of one energy quantum. 

 Moreover, he argues that it seems impos- 

 sible to explain the phenomenon in terms 

 of Bohr's one electron. He concludes that 

 a number of electrons must take part in 

 the emission of a single line, each having 

 the same frequency under ordinary condi- 

 tions or in a magnetic field, but different 

 frequencies when displaced unsymmetric- 

 ally in an electric field. It is difficult, 

 however, to understand why hydrogen has 

 only one detachable electron if Stark's 

 view is correct. 



It has already been mentioned that at 

 low pressures the width of lines may be 

 ascribed entirely to the Doppler effect. 

 The great broadening at higher pressures 

 has never been explained, but it has been 

 assumed that damping, collisions and ro- 

 tations all play a part. Stark suggests that 



it may be largely due to atomic electric 

 fields, which may exercise a large influence 

 when the atoms are crowded together. It 

 seems significant that the broadening in- 

 creases with the ordinal number of a line 

 in a series, is often unsymmetrical, and 

 diminishes with increasing atomic weight 

 in most cases, quite in harmony with the 

 effects of an electric field. Nicholson and 

 Merton have found that the broadening of 

 hydrogen lines is in quantitative agree- 

 ment with Stark's suggestion. 



With changes in vapor density, pres- 

 sure, temperature or the mode of excita- 

 tion lines belonging to one series may 

 weaken or disappear, other lines may be 

 strengthened, and new lines may appear. 

 We must assume that different groups of 

 lines are due to different emission centers. 

 These differences must depend upon the 

 size of the particles, or upon the number 

 and arrangement of electrons. Any theory 

 must take account of the molecular or 

 atomic state or the electrical charge of the 

 emission centers. In some cases we have 

 rather definite information on these points. 



A number of elements emit band spectra 

 under some conditions, line spectra under 

 others. One conclusion which seems to be 

 well established is that band spectra are 

 emitted by molecules, line spectra by 

 atoms. Universally we find that com- 

 pounds give band spectra, never line 

 spectra. If a compound is dissociated by 

 the discharge the line spectrum of one or 

 both constituents appears. Elements give 

 band spectra with feeble excitation, line 

 spectra when the discharge is so intense as 

 to cause dissociation. It seems reasonable 

 to infer that the band spectra of elements 

 is likewise associated with the molecular 

 condition. In the case of monatomic ele- 

 ments which give both band and line spec- 

 tra electrical conditions must determine 

 the nature of the radiation. 



Radiation is an electromagnetic process, 



