EXPERIMENTS WITH THE DISPLACEMENT INTERFEROMETER. 93 



i' f 2', 3', to enter the telescope. Similarly the transmitted rays 4, 5, 6, after 

 reflection from the opaque mirror M, are separated into the three groups 

 4', 5', 6' reflected from the front, the intermediate, and the rear faces of the 

 grating towards the telescope. It follows, therefore, that the ray 6', reflected 

 from the rear face only, can have no spectrum, since it does not pass through 

 the film g. The others, i, 2, 3, 4, 5, do pass through the film, and when super- 

 posed in pairs must give rise to elliptic interference in the superposition of 

 the spectra. The figure shows the refracted rays R and the diffracted rays D, 

 just before entering the telescope. 



Since the glass is ordinary plate, it is almost invariably slightly wedge- 

 shaped and consequently direct rays will be seen in the telescope as six parallel 

 lines. These are shown in fig. 53 in the form actually observed with the present 

 grating. The rays i, 4 from the front, 2, 6 from the rear of the plates, are 

 dazzlingly white, whereas 3 and 5 from the film within are respectively 

 brownish-yellow and yellowish. The latter are thus easily distinguished, but 

 are otherwise adequately white and perfectly sharp. It follows also that the 

 air-film is plane parallel, otherwise there would be two images from this re- 

 gion; that it does exist, however, is shown by the reticulated interferences 

 below, 55 and 56. 



If the group of rays i, 2, 3, is separated, all but one may be screened off, 

 and this is frequently possible. It is much more difficult, however, to screen 

 off the rays 4', 5', 6' individually as a rule, though if the angle of incidence is 

 large (45 as compared with 15 in this paper) this may also be done. No. 6' 

 is of no consequence, since it does not appear in the spectrum; No. 4' may 

 be removed by blacking a vertical line at the point a in the diagram; or a 

 narrow screen (vertical rod) may be placed at the proper position just behind 

 the grating. The object in view is to admit only the two spectra which are 

 to interfere in the telescope, as this sharpens the black lines enormously. 



It is for this reason that it is undesirable to have the rays return upon 

 themselves; i.e., they should not be reflected from M and TV quite normally. 

 In such a case, as will presently appear more clearly, the blotting out of indi- 

 vidual rays or spectra is more easily accomplished. 



In order that the slit images may be superposed horizontally and vertically, 

 a fine wire is drawn across the slit of the collimator, the wire being imaged by 

 the black spot on each line, as suggested in fig. 53. The adjustment screws 

 (horizontal and vertical axes of rotation) on the mirror then enable the observer 

 to bring any two slit images and the corresponding spots into coincidence. 

 This adjustment must be made accurately, if the interferences are to be seen 

 in the spectra in the field of the other telescope. 



54. Continued. The groups 1+4, 1+5. The character of these interfer- 

 ences, in which the rays i and 4 or i and 5 are superposed horizontally and 

 vertically, is shown in figs. 54 and 5 5 . The rays do not return upon themselves, 

 L being the impinging white vertical sheet of light, M and N the opaque 

 mirrors, g the film inclosed between the plates GG' which are equally thick. 



