412 Prof. Carl Bams on Interferometry 



the two white and the two yellow images of the slit be put 

 in coincidence and the mirrors M and N are adjusted for 

 the respective reduced or virtual path difference zero, the 

 interferences obtained are usually eccentric ; i. e. the centres 

 of the interference ellipses are not in the field of view. The 

 effective reflexion in each of these cases takes place from 

 the front and rear face of the grating at the same time. 

 Hence the interference pattern includes the prism angle of 

 the grating plate and is not centred. The air-paths of the 

 component rays are here practically equal. In addition to 

 the ellipses, this self-compensating position also shows 

 revolving linear interferences, and (as a rule) a double set 

 is in the field at once, consisting of equidistant symmetrically 

 oblique crossed lines, passing through horizontally in 

 opposite directions together, when either mirror M or N is 

 suitably displaced. 



If either pair of the white and yellowish images of the 

 slit be placed in coincidence when looking along EG, the 

 interference pattern along DG is ring-shaped, usually quasi- 

 elliptic and centred. The light returned by M and N is in 

 this case reflected from the same face of the grating, either 

 from the face carrying the grating or the other (unruled) 

 face. The corresponding air-paths of the rays are in this 

 case quite unequal, because the short air-path is compensated 

 by the path of the rays within the glass plate. Hence these 

 adjustments are very different, in one instance GM, in the 

 other GN, being the long path. For the same motion of 

 the micrometer-screw, the fringes as a whole are displaced 

 in opposite directions. In one adjustment there may be a 

 single family of ellipses ; in the other there may be two or 

 even three families, nearly in the field at once. 



If the grating were cut on optical plate glass, the adjust- 

 ment for equal air-path would probably be best. But with 

 the grating cut as usual on ordinary plate, or in case of 

 replica gratings on collodium or celluloid films, the adjust- 

 ment for unequal paths is preferable. Here, again, one of 

 the positions is much to be preferred to the other, owing to 

 the occurrence of multiple slit images from one of the 

 mirrors, as above specified. In fig. 2, for instance, where 

 the grating face is to the rear, there are but two images, 

 1 and 2, from M if the plate is slightly wedge-shaped ; but 

 from N, in addition to these two normal cases (not necessarily 

 coinciding with 1 and 2), there are two other images, 3' and 

 4', if 3 and 4 are spectrum rays, resulting from double 

 diffraction, with a deviation 6, and angle of incidence I, 

 respectively 6 < I and 6 > I, in succession ; or the reverse. 



