REVERSED AND NON-REVERSED SPECTRA. 



39 



by the telescope, the distance e\e\ may be too large to admit of appreciable 

 interference. Hence the colored strip within which interferences occur will 

 comprise those wave-lengths which lie very near X, whereas the colors lying 

 near X', etc., will be free from interference. 



If, however, the mirror M is displaced parallel to itself to M' by the microm- 

 eter-screw, the rays c"d" and c'd' will now coincide at e'i, whereas the rays 

 from ab and a'b' will no longer issue coincidently and may not interfere. 

 Thus the interferences are transferred as a group from rays lying near X to 

 rays lying near X'. It is obvious, therefore, that with the displacement of M 

 the strip carrying interferences will shift through the spectrum and that an 

 enormous play of the micrometer-slide at M will be available without the loss 



25 



26 



..t 



r 



of interferences. In fact, a displacement e of over 3 cm. of M normal to itself 

 produced no appreciable change in the size or form of fringes, but they passed 

 from the green region into the red. In consequence of the film gratings used, 

 the strip in question was naturally sinuous and somewhat irregular, but the 

 fringes themselves in the clear parts were straight, parallel, strong lines. They 

 did not thin out to hair-lines at their ends, nor show curvature, as one would 

 be inclined to anticipate. On the contrary, they terminated rather abruptly 

 at the edges of a strip occupying about one-fourth of the visible length of 

 the spectrum. 



It follows, therefore, from figure 25, that the displacement of Mdoes not 

 change path-length or path-difference, for the rays are inclosed between 

 parallel planes, as it were. Since the double angle of reflection is 5 = i 80 2 6, 

 where 6 is the angle of diffraction of G and G', the displacements of M 

 over a normal distance e will shorten the path of M in accordance with the 

 equation 

 (i) n\ = ie cos 5/2 = 2^ sin 



