1010 Prof. R. W. Wood and M. G. Ribaud on the 



innumerable absorption lines in the yellow, green, and blue 

 region of the spectrum. By means of a concave grating of 

 11 feet radius, the spectrum of the transmitted light was 

 resolved into bright lines, equalling in narrowness the 

 emission lines of the iron arc. Similar results were obtained 

 with the vapour of sodium * and bromine f, and the spectra 

 obtained in this way were named magnetic rotation spectra. 



In the case of the vapour of sodium it was found that in 

 the red and orange region, some of the absorption lines 

 rotated the plane of polarization to the right, others to the 

 left. This phenomenon was observed by employing a double 

 prism of Fresnel (right and left handed quartz), which, when 

 employed in the well-known manner of Macaluso and Corbino, 

 causes the appearance of horizontal dark bands in the 

 spectrum. 



Selective rotation of the medium manifests 'itself by the 

 penetration of light from the bright into the dark bands. 

 In the red and orange region, bright needles of light were 

 observed to shoot into the dark region as soon as the magnet 

 was excited, some of them projecting themselves downward, 

 others upward, indicating positive and negative rotations of 

 the plane of polarization. 



In the green and blue regions of the spectrum, the rotations, 

 while sufficient to give a brilliant bright line spectrum, 

 were insufficient to make observations with the double prism 

 possible. It must be remembered that a rotation of 90° is 

 necessary to cause the light to pass from the centre of a 

 bright band to the centre of a dark one. In the case of both 

 sodium and iodine comparatively few of the absorption lines, 

 of which there are many thousand, appeared to rotate the 

 plane of polarization to an appreciable degree, the magnetic 

 rotation spectrum being made up of something over 100 lines 

 altogether. In the case of bromine, observations made with 

 a concave grating of 1'60 m. focal length (when used with a 

 collimating lens) have shown bright lines (rotation lines) 

 for all of the absorption lines which the grating was capable 

 of resolving, in other words the absorption spectrum 

 and the magnetic rotation spectrum were complementary. 

 This was, however, only true when the vapour was at very 

 low density. At higher densities the appearances were 

 totally different. 



Recent work on the resonance spectra of iodine has 

 shown that a resolving power of at least 300,000 is necessary 

 for an exact study of the phenomena produced by all of 



* R. W. Wood, Phil. Mag. xii. p. 499 (1906). 

 t G. Ribaud, C. B. civ. p. 900 (1912). 



