Separation of the Spectral Lines of Thorium 3 



could pass readily from one line to another coincidentally upon 

 both sets of plates. In this way a series of readings was then 

 taken for a number of lines in the field of view upon both sets of 

 plates. The majority of these adjacent lines were duplicates in 

 position within the errors of setting the microscope. This error 

 for the ^-components was twice as great as the corresponding 

 error for the no-field lines. The latter also can be adjusted more 

 accurately than the double readings of the ^-components of a 

 triplet. 



By this method of comparison I discovered that a great many 

 of the lines were unsymmetrical in position, assuming that the 

 lines which could be brought into juxtaposition were symmetrical. 

 If, however, as Professor Voigt writes me, there should be a one- 

 sided displacement of the p- or ^-components, it might be impos- 

 sible to detect it, as the plates would be displaced as a whole. 

 Such a displacement or dissymmetry is theoretically derived by 

 Professor Voigt. 1 It involves the period of the vibrating ion and 

 a constant which depends upon the absorption and emission of the 

 ion. This constant would vary from ion to ion, and therefore 

 there would be large variations from line to line, unless the lines 

 arose from the same ion. From Geiger's 2 investigation, the dis- 

 placement is only one two-thousandth part of the total separation 

 of the D 1 line for a field strength of 2,400 lines to the square cen- 

 timeter. But for the red K-Yme corresponding to D. 2 Geiger's 

 data give a ratio of 3 per cent. The latter would be observ- 

 able unless the lines were diffuse. We may, then, in a great many 

 cases expect the effect to be too small to be observable. Even if 

 it were generaly observable in any particular substance, the great 

 range in values would make it impossible to set the greater part 

 of the lines in juxtaposition on the two sets of plates at the same 

 time. Therefore, the juxtaposition of most lines indicates that 

 the effect is too small to be observable in those cases. For the 

 weak lines, the effect being a function of the emission, the dis- 

 placement should be small, and probably smaller than I could 

 detect. On account of the width of many of the ^-components, 



*Phys. Zeit. 9 (190S), p. 123, eq. 11. 

 2 Ann. d. Phys. 24 (1907), p. 597. 



91 



