190 



F. E. Wright — Transmission of Light through 



clockwise); Id for the position +i, + &>; 10 for the position 

 + /, 180° —&) ; 10 for the position -ti, 180° + <b. In averaging 

 the readings, all these positions should, with the exception of 

 the sign, furnish identical results. This was fonnd to he the 

 case for all positions of the plate for which good positions of 

 darkness were attainable. Occasionally, however, differences 

 of 20' or even 30' between the positions were obtained, even 

 on good plates. The observations were made on bare cleavage 

 plates, 1 to 3""" thick, on which the two refracted waves from 

 a single incident wave were appreciably separated after trans- 

 mission. A large area of the front surface of each plate was 

 blackened by means of a dull lacquer and rendered opaque 

 (shaded part of fig. 8a). When viewed through the plate after 



Fig. 8a. 



Fig. 86. 



rotation through 180° about the horizontal axis, the exposed 

 area of fig. 8a appeared as in 8b. Two images were observed, 

 one from the ordinary wave O and the second from the extra- 

 ordinary, E. These two images overlapped to a large extent, 

 but along the margins of the bright part of the plate the image 

 of the extraordinary wave E extended on the left beyond that 

 of the ordinary wave, while on the right and at the top a nar- 

 row strip, due entirely to the ordinary wave O, appeared as 

 indicated in fig. 8b. The azimuth of the plane of polarization 

 of one of these emergent waves, E or O, was then ascertained 

 by rotating the microscopic stage Hj until the margin in ques- 

 tion disappeared completely. It was found that within the 

 limits of error each edge became completely dark (1) when the 

 azimuth of the plane of polarization of the entering waves 

 was such that the amplitude of its refracted wave was zero 

 (uniradial azimuth of second entering wave) ; and (2) when the 

 plane of polarization of the emergent wave coincided pre- 

 cisely with the extinguishing plane of the analyzer. These 

 two positions were, within the limits of error, precisely 90° 

 apart, thus proving experimentally the relation (37) deduced 

 from theory ; in other words, the uniradial azimuth of an 

 entering wave W, is 90° from that of the second emergent 

 wave, W' a . The doubly shaded portion of fig. 8b is the area 

 over which the two images overlap. The positions of extinc- 



