ON LIGHT. 353 



sitely polarized, the one being reflected, the other trans- 

 mitted, and intermingled with the unpolarized part. 



(134.) If a parallel plate of glass be used for this ex- 

 periment, the same process is repeated at the hinder 

 surface. An equal per-centage of the unpolarized por- 

 tion is similarly divided between the reflected and 

 transmitted rays, oppositely polarized, and as the trans- 

 mitted polarized portion is, ipso facto, guaranteed from 

 subsequent reflection at the polarizing angle, the total 

 amount of polarized light in each of the two beams is 

 nearly doubled. If behind this a second parallel glass 

 plate be applied, the same process is again repeated on 

 the remaining unpolarized portion, and so, by multiply- 

 ing the plates, the whole incident beam is ultimately 

 divided equally into a reflected and a refracted beam 

 completely polarized in opposite planes. Such at least 

 would be the case were the plates perfectly transparent 

 and infinite in number ; but as these conditions cannot 

 be fully realized, the transmitted beam is never com- 

 pletely polarized, though enough so to afford a very 

 convenient mode of viewing many optical phsenomena. 

 On the other hand, if the plates be truly plane and their 

 surfaces exactly parallel, the reflected beam is wholly 

 polarized, and as its intensity is nearly half that of the 

 incident light, this affords an excellent mode of procur- 

 ing a polarized beam available for purposes of experi- 

 ment. A frame containing six or eight squares of good 

 window glass laid one on the other, and backed by a 

 sheet of black velvet, is one of the most convenient and 

 useful of optical instruments, and will be frequently 



