Interference with Reflecting Gratings. 121 



with the E line is 20', their difference %i, and the rays D v 

 R, D' intersect at a common centre on G m . Hence, if we 

 place the plane of G m at the centre of the spectrometer and 

 arrange M and Gr„ eccentrically, the angles may be measured 

 as before. 



3. Resolution of the Slit Image. — If the sharp white images 

 of the slit in a Michelson apparatus for the case in which the 

 incident light consists of parallel white rays from a colli- 

 mator, be accurately superimposed, and the opaque mirrors 

 be set at the proper distances from the semi-transparent 

 mirror by the micrometer, the slit image may itself be viewed 

 through a grating and will then show elliptic interferences 

 in all the spectra. The apparatus is here eccentric, while 

 the grating (either transmitting or reflecting) must be at the 

 centre of the spectrometer, if angles are to be measured.. 

 The same is true for any of the other superimposed white 

 slit images in the above or the earlier experiments, and may 

 even be repeated with successive transmitting gratings. It 

 is interesting to note that the position of the centre of ellipses, 

 is at the same wave-length in all the spectra, though the form 

 of ellipses may differ enormously. The same phenomenon 

 may thus be seen by a number of observers at the same time r 

 each looking through his own telescope. 



4. Third Method. Parallel Gratings. — In this case the 

 two halves of the grating treated are displaced parallel to 

 themselves, from their original co-planar position in the 

 grating, from which they are cut. Their mounting is thus 

 something like the case of the two black plates of Fresnel's 

 mirror apparatus, one of the plates being adapted for dis- 

 placement parallel to itself. 



Fier. 2. 



In fig. 2, g and g' show the two halves of the reflecting 

 grating, cut along the plane S, normal to the plates and 



