172 TRANSACTIONS OF THE [apR. 1, 
tions of the field as appeared to have a uniform illumination, and 
it is not thought that the error due to this cause could have been 
great. 
Phosphorescent bodies were also examined for polarization, but 
the light emitted by such bodies is so weak that no definite re- 
sults were obtained. 
B & 
APPLICATION OF FRESNEL’S FORMUL FOR VITREOUS REFLECTION. 
The main object of this research being to determine whether 
or not polarization by emission could be experimentally proven 
to bea phenomenon of refraction, Fresnel’s laws for reflection 
and refraction, which have been shown by many experiments to 
accurately represent the facts, were now applied to the 
determination of the amounts of polarization which should 
be produced by single refraction of light passing through the 
boundary surface between uranium glass and air. In order 
to apply these laws it is necessary to assume that all of the 
light emitted by the uranium glass, whether coming from the 
surface molecules or from the interior layers, has undergone 
the process of refraction—an assumption not contained in 
Arago’s explanation of the cause of the phenomenon. 
Taking the intensity of the incident ray as unity, Fresnel’s 
formule give for the intensities of the reflected and refracted 
rays, when the incident beam is plane polarized in the plane of 
incidence, 
sin? (a — 3) 
reflected ray =r, = ane(eoR) 
5 Sry: 
4 cos? a sin? 3 
sin? (a + /3) 
refracted ray =7,' = 
a being the angle of incidence and £ the angle of refraction. 
For a ray plane polarized in a plane perpendicular to the 
plane of incidence 
tan2 (a — 8) 
reflected ray = 7, = fant ea Bh 
rey Shite: 
4 cos? a sin? 3 
~~ gin? (a + /3). cos? (a—/) 
refracted ray =7,/ 
Since ordinary light may be considered as composed of two 
equal plane polarized beams, polarized in planes at right angles 
