1012 Prof. R. W. Wood and M. G. Ribaud on the 



It was hoped that observations could be made with the 

 Fresnel double prism, but preliminary experiments showed 

 that the rotations were too small to cause any appreciable 

 penetration of light into the region of the dark bands. It 

 was necessary, therefore, to make use of the method 

 previously used in the study of the magnetic rotation of 

 sodium in the green region of the spectrum. The polarizing 

 nicol is rotated until sufficient light is restored to render the 

 absorption lines visible (10 to 15 degrees). The magnet 

 is then excited and the spectrum brightens at the points where 

 the rotation is in the opposite direction to' that in which the 

 nicol has been rotated, and darkens where the rotation is in 

 the same direction. 



Suppose that the rotation is positive to right and left of 

 one absorption line, and negative to the right and left of 

 another. The former will appear narrower when the field is 

 excited as a result of the brightening of the edges of the 

 dark line. The latter will, however, appear broader than 

 in the absence of the magnetic field. If the rotation is 

 anomalous, the centre of the absorption line will appear 

 slightly shifted as a result of its becoming brighter on one 

 side and darker on the other. This method of observation 

 gives results quite as conclusive as those obtained with the 

 Fresnel prisms, and is well adapted to cases where the 

 maximum rotation is less than forty or fifty degrees. 



It was found necessary to orient the analysing nicol so as 

 to obtain the maximum reflexion from the grating, for the 

 polarizing power of a grating is very large, especially in 

 spectra of higher orders than the second. In the present 

 instance the difference in the reflecting power was certainly 

 five or six fold for vibrations parallel and perpendicular to the 

 oTooves. With this arrangement of the apparatus we had 

 no difficulty in observing rotations both positive and 

 negative in direction, but the intensity of light was not 

 quite sufficient to enable us to be sure of what happened on 

 both sides of the lines. The rotation appeared to be more 

 marked on one side of the lines than on the other, and we 

 felt uncertain about our results. 



We accordingly substituted for the grating a very fine 

 echelon, loaned through the kindness of Mr. I\ Twyman, 

 of the Hilger Company. This instrument consisted of 

 20 plates in optical contact (each plate 15 mm. thick), and 

 gave a resolution equal or superior to that of the grating 

 and an image of much greater intensity. 



It was immediately obvious, with this instrument, that 

 certain lines became broader when the magnet was excited, 



