of Emission and Absorption Lines in a Gas-Spectrum. 285 



intensity. One of these lines would be invisible in a small 

 magnetic field, but in a strong magnetic field the two lines 

 would be of equal intensity. Moreover, the separation pro- 

 duced by the magnetic field is superposed in the manner 



above indicated on an initial distance \-^) apart oE the 



two lines ; this distance is precisely that calculated as the 

 pressure shift of the single line, and so we can assert that 

 those lines which show a good pressure shift will also show 

 a good separation of the lines under the action of a magnetic 

 field, as the separation always starts with this initial value. 

 Similar conclusions are obtained by considering the other 

 components of the wave oscillation (Voigt's S-wave). I 

 have worked this case out, but the results exhibit nothing 

 new, and I shall not repeat them here. 



II. Theory for Emission Spectrum. 



I now propose to construct a theory for the emission 

 spectrum corresponding to the absorption theory just deve- 

 loped. Experiment indicates that the emission and absorption 

 lines are exactly similarly affected in this respect. It will, 

 however, first be necessary to discuss the subject of the 

 emission of the light disturbance by a gas rather more 

 fully. 



As we have previously shown, the subject can be discussed 

 in terms of the real negative electrons, neglecting altogether 

 the positive charge and its motion. There are of course 

 certain effects which can be attributed to these two causes, 

 such, for instance, are the phenomena of band-spectra and a 

 definite width of the lines in line spectra. These matters are, 

 however, independent of the subject under consideration, and 

 may be left out of account in the discussion. 



We thus confine the discussion to the negative electrons, 

 typified by the electron with a charge e, whose displacement 

 from its position of equilibrium in the atom is determined by 

 the vector whose components are (#, y, z) referred to fixed 

 directions in space. 



Now consider the emitting gas (or vapour) as a whole, 

 and take any volume element dr of it, which contains a 

 crowd of molecules numerous enough to be expressible con- 

 tinuously. We can investigate the radiation from this 

 element in terms of the oscillating electrons in the molecules 

 in it. 



Now Larmor has shown that there is no radiation from 



