﻿Magnetic Rotation Spectra of Sodium Vapour. 511 



These wave-lengths I consider to be accurate to within 

 about 2 A.E. or \ of the distance between the D lines. 



It is, of course, of the utmost importance to determine the 

 law which governs the spacing of the lines in the simple 

 spectra. A criterion may perhaps be obtained by referring 

 to the magnetic rotation spectrum of the vapour, the lines 

 of which correspond in general to the lines of the fluorescence 

 spectrum. This spectrum has been photographed with a 

 large concave grating and the wave-lengths determined 

 certainly to within a tenth of an Angstrom unit. The strong 

 lines at the following points of the spectrum form a series 

 analogous to the series obtained with the cadmium 480 

 series. 



X. X differences. 



SJ?5J 39-56 A.E. 



5079-78 oq.-i o 



5040-G5 on.os 



522S :::::::::: 2m 



4962-85 qq.co 



4924-32 * W 



This we have called the first series. 



The wave-length differences are, in this case, much more 

 nearly constant, and decrease progressively. 



The lines of this series are especially conspicuous in the 

 magnetic rotation spectrum (PL XL, fig. 1, c), hence I have 

 mentioned it first; they appear in the fluorescence excited 

 by the lead line 5001, as will be seen by reference to the 

 chart. 



If now we try to fit one of the magnetic series to the 

 cadmium 480 fluorescence series, we find that the third 

 magnetic series coincides with it between 5019 and 5134, 

 while in the violet region it coincides with the fourth maff- 

 netic set. I do not feel sure whether this peculiarity is due 

 to slight errors in the determination of wave-lengths or not. 

 I think not, however, for I have very carefully superposed 

 the two negatives (cadmium fluorescence and magnetic 

 rotation spectra), both taken with the same instrument, and 

 find the same disagreement. We cannot be sure of anything, 

 however, until the cadmium series has been photographed 

 with the grating. 



During the coming year I expect to photograph the 

 fluorescence spectrum excited by cadmium and zinc radiations 

 with the large concave grating. It will then be possible to 

 determine the wave-lengths of the lines to within a tenth of 

 a unit. One of the most remarkable facts connected with 

 the appearance of the lines of a series is that the distribution 



