CHAPTER VII. 



THE SELF-ADJUSTING INTERFEROMETER IN RELATION TO THE ACHROMATIC 



FRINGES AND REFRACTION. 



54. Introductory. This apparatus was first used for the case of coincident 

 ray systems by Michelson and Morley * in their famous experiments on the 

 Fresnel-Fizeau coefficient and has since been similarly employed by Zeeman.f 

 It has so many practical advantages that a special reference here is justified. 

 It is shown in an essentially modified form in figure 79,$ where L is a pencil 

 of white light, preferably from a collimator. It is separated by the half- 

 silvered plate N into the two beams Li234$T and LidySjT, both of which 

 are recombined at 5 and then enter the telescope at T. It is merely necessary 

 to rotate any of the mirrors, say N', around a vertical axis until the two ver- 

 tical white, wide slit-images coincide in the telescope, when brilliant fringes 

 will be at once obtained on the coincident white fields. The central fringe is 

 achromatic, for the system is self -compensating, or the glass-paths are rigor- 

 ously equal. The fringes may be enlarged to infinite size and then reduced in 

 size again, the phenomenon being symmetrical, by 

 rotating any mirror, say N', about a horizontal axis. 



The mirrors N and N' are rigidly fixed to a carriage 

 capable of sliding right and left parallel to the lines 8, 

 5, etc. Hence the rays 8, 5, and 2, i and the rays 3, 4 

 and 7, 6 may be brought to coincidence (cf . fig. 94) 

 or separated in any degree at pleasure. If the slide 

 were perfect the fringes would not be disturbed by 

 this process; but few are perfect to this degree, and the fringes will change 

 size somewhat, since there is rotation. Practically this is of no consequence. 



The fringes, which when sharp are necessarily horizontal, may also be 

 changed in size by inserting a plate-glass compensator, C, about 5 mm. thick, 

 in the paired beams 8 5, 2 I, or 3 4, 76. When this is rotated on a horizontal 

 axis the fringes pass through infinite size, and this arrangement is particularly 

 adapted for the detection of the character of the fringes and will be so used 

 presently. 



If a direct- vision prism or grating is placed in front of the telescope, the 

 spectrum is seen to be crossed by intense black lines, very nearly parallel and 

 horizontal, but actually diverging from blue to red symmetrically up and 

 down from the horizontal central black line. It is not necessary here, that 

 the slit be fine. In fact, it may be several millimeters broad without destroy- 

 ing these spectrum fringes, if essentially horizontal. 



* Amer. Jour. Sci., xxxi, p. 377, 1886. 



f Proceedings Amsterdam Acad., vol. xvn, 1914, p. 445; and vol. xvin, 1915, p. i. 



t The Michelson-Morley apparatus does not admit of appreciable ray separation, for 

 on displacing any of the mirrors, the rays entering the telescope separate. This is not the 

 case in figure 79. 



74 



