Mr. R. T. Grlazebrook on NicoVs Prism. 249 



while the plane of polarization of the light in the prism is no 

 longer fixed relative to the prism. 



Consider a sphere, centre 0. Let ON be parallel to P Q, 

 the intersection of the wave-fronts and the face EFGH; 

 Y'j V parallel to the directions of vibration in the two 

 waves. Then NOV^NOV' are parallel to the wave-fronts, 



Also since 



cos (f> — (f> / =cotO tan 6 f , 



N V V is a right angle, and 



sin (^ tan VV 7 cot (<£-(//)• 



V V is the angle between the planes of polarization =% (say) ; 

 then 



tan^= sin 0' tan (<£ — <£') (2) 



6 f is the angle between P Q and the possible direction of vibra- 

 tion in the crystal corresponding to a wave cutting EFGH 

 in P Q. 



To find this direction, draw a plane through B D perpendi- 

 cular to the optic axis ; the line in which it cuts the wave- 

 front in the crystal either coincides with or is perpendicular 

 to the direction of vibration. Take 

 the two as coincident. As before, let 

 N be parallel to P Q, V' to the 

 direction of vibration, L to the line 

 F H. Then N L is the face of inci- 

 dence, NOV 7 the wave-front inside, 

 L V the plane perpendicular to the 

 optic axis. Also 



Let 



KL = f, NLV' = «. 

 a is known, being the angle between a plane perpendicular to 

 the optic axis and one of the rhombic faces. We can show 

 that 



a = 45° approximately; 



yfr depends on the direction of the incident wave, and can be 

 found when that is given. And we have 



cot 6' sin yjr = cos yjr cos $ + sin <// cot a. . . (3) 



We must therefore determine -ty. 



Take, as axes of ,v, y, z respectively, a normal to ABCD 

 Phil. Mag. S. 5. Vol. 10. No. 62. Oct. 1880. T 



