686 Prof. C. V. Raman and Mr. K. Seshaoiri Kao on 



e> 



This corresponds to light polarized in the plane of 

 incidence. For light polarized in a perpendicular plane, 



2tt, _ Qq 2 + 1) -Q^ + l) cos 2 (r-i ) 



These two formulae should give slightly different values 

 for the positions of the rings. That such a difference 

 actually exists is verified on observation of the rings of 

 non-uniform plates obliquely through a nicol, when it will 

 be found that the rings shift slightly on rotating the nicol. 

 The effect is, however, perceptible only for fairly large 

 angles of diffraction. For instance, numerical calculation 

 shows that the difference between the two values of 8 

 corresponds to an alteration of the path-difference of the 

 interfering rays of only 0*02 X for an angle of diffraction 

 9° 30'. For a deviation of 62° it increases to 014A,and 

 this gives an easily measurable difference in the position 

 of the rings, especially with films which are fairly thin. 

 This indication of theory has been confirmed by quantitative 

 measurement. 



The angles of diffraction at which the rings on the film 

 appear blurred may be readily calculated. As these angles 

 are generally small, we may use for this purpose an approxi- 

 mate form of formula (1) giving the path-difference directly 

 in terms of the angle of diffraction : 



O-l)£-4*sin0 = nX (3) 



Formula (2) may be written as 



({M-l)t + itsm0 = nX (4) 



The diffraction rings whose positions are given by (3) 



and (4) would be completely out of step if t sin 6 = , -= , 



etc. Table III. shows for comparison the calculated and 

 observed values of the angles of the first blurring in a few 

 cases exhibiting satisfactory agreement. The second blurring, 

 which occurs at a larger angle of diffraction, is hardly so 

 conspicuous as the first. 



Table III. 



(/x — l)t. Observed angle. Calculated angle. 



7 A 1° 30' 1° 20' 



3\... 3 40 3 20 



2 A 5 15 5 2 



