between two Substances of different refractive Powers, 27 



stage except a total absence of rings. From the agreement of 

 this with theory I conclude that the polarization of light at 

 the inner surface of glass is (to the senses) complete. But at 

 the polarizing angle of the diamond the case is perfectly dif- 

 ferent On increasing the angle of incidence till it exceeds 

 this angle, the white-centred rings do not disappear, but the 

 first black ring contracts so as to leave no central white, and 

 becomes itself the black centre. After this there is no mate- 

 rial change : I find, however, that the black centre of the rings 

 produced by light polarized perpendicular to the plane of re- 

 flexion is always (beyond tbe polarizing angle of the diamond) 

 sensibly larger than the black centre of the rings produced by 

 light polarized in the plane of reflexion. 



The nature of this transition from rings of one character to 

 rings of the opposite character appears to me to be, theoreti- 

 cally, extremely curious. As the rings do not disappear, it is 

 plain that if light polarized perpendicular to the plane of inci- 

 dence (or whose vibrations are entirely in that plane) is inci- 

 dent at what is called the maximum polarizing angle of the 

 diamond, a portion of it is still reflected. Still, however, on 

 increasing the angle of incidence the character of the rings is 

 changed : and this takes place at an angle where (so far as we 

 are entitled to conclude) there is nothing peculiar in the re- 

 flexion from the glass ; and we are therefore compelled to 



2 w 

 admit, that the incident vibration being a . sin — — (v t — x) 9 



when the angle of incidence is increased so as to exceed that 



2 it 

 angle, the reflected vibration is changed from + p . sin — 



A 



2 it 

 (v t — x) to — q . sin — (vt — x). A similar change takes 



A 



place at the polarizing angle of the glass : but there, as we 

 have seen, the transition from -f p to — q is effected by pass- 

 ing through 0, or by the entire cessation of reflexion at one 

 angle of incidence ; which is not the case at the polarizing 

 angle of the diamond. How then is the gradual change from 



2 W 2 7T 



+ p sin — (v t — X s ) to — q . sin — (vt— x) to be explained? 



u ,, . , A A 



I answer that the phaenomena prove that it follows from a 

 gradual change of phase, while the coefficient is not much 

 altered. In other words (neglecting the trifling alteration in 



the coefficient) the quantity -f p sin — (v t — x) is changed 



A 



2 "K 



to — p sin — (vt — x) % not by the disappearance of p, but by 



E2 



