150 Prof. E. W. Wood on the Magnetic 



For the verification of a formula for the magnetic rotation 

 of an absorbing medium, the results found for a rotation of 

 720° at the centre should be used, as this curve is probably 

 a little more accurate than the others. 



The curves A, B, G, D,E, and F are plotted from measure- 

 ments made on the photographic plates. Curve C, on the red 

 side of the D lines, lies a little too near Curve B. The error 

 was probably caused by an accidental slipping of the plate on 

 the dividing-engine, and was not discovered until the results 

 were plotted. Curves G-M are for rotations of 810°-1250° 

 at the centre. They are from the best of the visual obser- 

 vations, and are added to give an idea of the rotatory power 

 of vapours of greater density. 



Theory of the Magnetic Rotation, 

 Becquerel's formula for the magneto-rotatory power of a 



medium o = GX- i — shows the relation between 8 the rotation 

 d\ 



and the wave-length A,. 



It was developed from the hypothesis that the electron 

 system, which caused dispersion, was set in rotation as a 

 whole by the magnetic field. 



If O be the period of revolution of a train of circularly 

 polarized waves rotating in the same direction as the mole- 

 cules, and a) that of the molecular configuration, a circularly 

 polarized wave-train of period O + co will travel in the 

 magnetized medium with the same velocity as a train of 

 period O in the unmagnetized medium. The magnetization 

 of the medium will therefore accelerate or retard a circularly 

 polarized wave-train according to the direction of revolution 

 of the vibration, the rotation of the plane of polarization 

 being the result of the unequal velocity of its two oppositely 

 circularly polarized components. The derivation of the for- 

 mula will be found in Larmor's ' JEther and Matter/ p. 352. 

 (In last line read "same ;; for "opposite," however.) 



Drude has developed two formulae in his Lehrhich der Optik, 

 one of which (Hypothesis of molecular currents) is obviously 

 unsuited to the present case, since it calls for rotations of 

 opposite sign on opposite sides of the absorption-band. The 

 other (Hypothesis of the Hall effect) represents fairly well 

 the rotatory dispersion of sodium vapour, as 1 showed in the 

 earlier paper, at least for the region not included between 

 the D lines. The complete formula 



1/A B\ 2 C\ 2 \ 



