292 Professor Zeeman [March 30, 



three motions with periods T,T-\-v, T — v,v being a vsmall quantity. 

 To each motion of the electrons there corresponds a himinous vibra- 

 tion, according to the electromagnetic theory of light. Observing 

 with a spectroscope we must, therefore, see each spectral Hue divided 

 into three hues ; each line becomes a triplet* 



I will show you a few examples of hues which are really divided 

 into three components, in accordance with Lorentz's theory. (Fig. 2, 

 iron ; Fig. 8, zinc ; Fig. 4, part of iron spectrum ; Fig. 5, part of 

 same spectrum.)t You will notice that each of the components 

 reruains very narrow ; it is not a hazy effect, but a very definite one. 

 This certainly would not be the case if all molecules did not behave 

 in the same manner, and if certain conditions of isotropy of the 

 molecules were not fulfilled, f 



The consideration of the model may illustrate some other points 

 which were foreseen by Lorentz's theory. 



Consider the light emitted at a right angle to the hues of force. 

 The three kinds of Hght seen in this direction are each due to vibra- 

 tions of one kind, and therefore polarised. We can, therefore, 

 extinguish the hght of the central component, or of the two external 

 components of the triplet by a Nicol. In one half of the sHde shown, 

 the external components are extinguished ; in the other half, the 

 central one. (Fig. 6.) 



So, for the first time, we were now able to get polarised radiations 

 from the molecules of a gas. All attempts to produce such simple 

 vibrations from gaseous molecules had hitherto failed. 



With some lines, the central component and the outer ones differ 

 niuch in intensity. If this be the case, the spectroscope can be 

 dispensed with entirely, and we may observe a partial polarisation of 

 the light emitted by the vapour in the field as found by Egoroff and 

 Georgiewsky. 



We shall now consider the light emitted in the direction of the 

 lines of force. (Fig. 7.) 



It is seen at once that each line must split up into two components. 

 Moreover, both lines must be circularly polarised, but in opposite 

 directions. With suitable arrangements, in one half of the field of 

 view the one, in the other the second component can now be 

 extinguished. I observed this circular ] olarisation for the first time 

 [ n the case of the sodium lines now shown. You see how complete 



* Zeeman. Verslagen Kon. Akademie v. Wetenschappen, Amsterdam, 

 Mei, Juni, October, 1897. Phil. Mag., July and September, 1897. 



t The photographs illustrating this lecture are, excepting the diagrams, 

 enlarged copies from negatives. The scale is different in the various cases. 

 The separation of the outer components is of the order of one-sixth of the dis- 

 tance of the sodium lines {the vertical lines in Fig. 12). No. 2 is a copy of one 

 of the first photographs I obtained. The author is indebted to Prof. Runge 

 for Nos. 21, 22, 24. The nonet is not distinctly shown in the latter repro- 

 duction. 



X Lorentz. A.unalen der Physik, Bd. 63, p. 278, 1897. 



