418 Prof. Carl Barns on Interferometry 



Part IT. — Direct Cases of Interference : 

 Diffraction Antecedent. 



4. Diffraction he fore Reflexion. — If in fig. 1 the diffracted 

 beams (spectra) be returned by the mirrors M' and W to be 

 reflected at the grating, interference must also be producible 

 along GrD. Ag;iin there will be three primary cases : if 

 reflexion takes place from both faces of the grating at once, 

 the air-paths must be nearly equal, the grating itself acting 

 as a compensator. The interference pattern is ring-shaped, 

 but, as usual, very eccentric. If the reflexion of both com- 

 ponent beams takes place from the same face of the grating, 

 the interference pattern is elliptical and centred, and the 

 air-paths are unequal. The case is then similar to the pre- 

 ceding §§ 1, 2 ; but there is no direct or normal image of 

 the slit, as M and.N are absent. In its place there may be 

 chromatic images of the slit (linear spectra) at C due to the 

 specified double diffraction of each component beam in a 

 positive and negative direction, successively. But these 

 chromatic images are nevertheless sharp enough to complete 

 the adjustments for interference by placing the slit images 

 in coincidence. It is not usually necessary to put the 

 spectrum lines into coincidence separately (both horizontally 

 and vertically), as was originally done, both spectra being- 

 observed. Again, along GE (fig. 1) approximately, there 

 must be two successive positive diffractions of each com- 

 ponent beam, which would correspond closely to the second 

 order of diffraction. The advantage of this adjustment lies 

 in the fact that there are but two slit images effectively 

 returned by M' and 1ST, and hence these interferences were 

 at first believed to be stronger and more isolated. As a 

 consequence I used this method in most of my early experi- 

 ments, before finding the equally good adjustment described 

 in the preceding sections. 



5. Elementary Theory. — To find the path differences fig. 3 

 may be consulted. the grating face is shown at gg, the 

 glass plate being e cm. in thickness below it, and n is the 

 normal to the grating. M and N are two op;ique mirrors, 

 each at an angle © to the face of the grating. Light is 

 incident on the right at an angle i nearly 45°. In both 

 figures the rays y m and y n (air-paths) diffracted at an angle © 

 in air, are reflected normally from the mirrors M and N 

 respectively, and issue toward p for interference. The rays 

 y pass through glass. Both figures also contain two com- 

 ponent rays diffracted at an angle 6 in air, where 6 — © = «, 



