296 Professor Zeeman [March 30, 



Analysing the light by means of a Rowland grating we can pro- 

 duce such a system of fringes for all wave-lengths, and we can consider 

 the rotation for wave-lengths close to the controlling absorption 

 bands. On the screen I will first project the fringes close to the 

 sodium lines with the field off. The dark vertical lines are the 

 sodium lines. They are broad, because the vapour is rather dense. 

 The horizontal bands are the interference fringes. (Fig. 12.) 



With the magnetic field on, the image now projected is seen. 

 (Fig. 18.) 



You see how fast the rotation increases in the vicinity of the 

 absorption lines, becoming more than 180° closer to the bands. In 

 the interior of the bands only a hazy fringe is seen. A remarkable 

 equation, first deduced by Becquerel,* gives the law of the rotation. 



The phenomenon is more beautiful as soon as the vapour is so 

 thin that the doublet is seen. (Fig. 14.) 



Outside the components of the doublet the fringe rises iqnvards. 

 But inside the components the fringe has moved doivmvards, the 

 rotation is negative there. The rotation is —90° for D^, nearly 

 — 180° for D.,. It is very interesting to watch the movement of the 

 fringes in the spectroscope as the field is increased or the density of 

 the vapour changed. 



DouhU Refraction and Resoliition of tlip Absorption Lines. 



In the second place we will now consider the double refraction 

 which occurs whenever hght traverses a vapour at right angles to the 

 magnetic field. A plane wave with vibrations parallel to the field 

 has a velocity different from that of a wave with vibrations at right 

 angles to the field. It is only close to tlie absorption band that the 

 difference becomes perceptible. Sodium vapour in a magnetic field 

 behaves as a double refracting crystal for light close to the sodium 

 lines. This result of Voigt's theory was verified hj him in conjunc- 

 tion with Wiecheit in the case of dense vapours, and commented upon 

 ])y Becquerel and Cotton. 



'With great density, and using the same system of interference 

 bands, the phenomenon assumes the appearance now projected. 

 Whereas the rotation of the plane of polarisation was symmetrical on 

 both sides of the absorption band, you see that the double refraction 

 is not. On one side of the absorption line, sodium vapour behaves 

 b'ke a positive crystal, on the other side like a negative one. 



With very dilute sodium vapour, and with a magnetic field strong 

 enough to resolve the sodium lines, the theory must be extended. 

 There is no difficulty here. 



The observations made by Mr. Geest, as well as by myself, con- 



* Becquerel. C.R. 125, p. 679, 1897. Cf. also Schuster. The Theory of 

 Optics, p. 291-294, 1904. Siertsema. Proc. Ak., Amsterdam, 12, p. 499, 11)03. 



