Origin of Spectra. 215 



other spectra, recorded in the paper, make their appearance 

 in turn. Reasons can be given for thinking that these more 

 complicated spectra are due to the formation of definite 

 •electronic systems, which are practically unstable new sub- 

 stances formed from the normal mercury atoms by the passage 

 of the discharge, and that they are not due merely to different 

 modes of vibration of the electrons in someone system, such, 

 for instance, as the mercury atom. 



In many respects mercury appears to be an ideal substance 

 to use for investigating the origin of spectra. Its vapour is 

 monatomic and can be rendered luminous in several ways : 

 by the electric arc, or spark, or by the vacuum-tube discharge. 

 This latter may be produced (a) by an induction-coil discharge 

 with variable capacity and self-induction in the circuit, (b) by 

 means of the low potential discharge from a glowing lime 

 cathode, (c) by electromagnetic induction in the electrodeless 

 ring discharge. Under these different conditions a continuous 

 spectrum, a band spectrum, and spectra with different numbers 

 of bright lines may be obtained. 



As the result of a careful investigation of the spectrum of 

 mercury, Stark * has concluded that mercury possesses two 

 distinct line spectra, that given by the arc in vacuo and that 

 given by a vacuum-tube discharge. These he identifies as 

 being due to monovalent and to divalent atoms — atoms which 

 have lost one and two electrons respectively. Stark's views 

 on the origin of spectra may briefly be summarized as 

 follows : 



(1) Continuous spectra are due to freely moving electrons, 

 the accelerations of which are continually being changed. 



(2) Line spectra are due to the vibrations of the electrons 

 contained in the positive ions, all the lines of a series being 

 due to the same source ; but a single positive ion may be the 

 source of several series of lines at the same time. 



(3) Band spectra are due to electrically neutral bodies, 

 and are formed in the act of combination of the negative 

 electron with the positive ion. 



In addition to the conclusion that mercury should have 

 two distinct line spectra, Stark also holds that there should 

 be two distinct band spectra, one formed by the combination 

 of a divalent positive ion with a negative electron to form a 

 monovalent ion, the other produced when the monovalent ion 

 unites with another electron to form a neutral atom. 



These views of Stark have been criticised by W. Wien f, 

 who, from observations of the effect of a magnetic field on 



* Stark, Ann. d. Phys. xvi. p. 490 (1905). 

 t Wien, Ann. d. Phys. xxx. p. 349 (1909). 



