510 



The Rev. S. Haughton on some new Laws 



cidence, the intensity of the reflected light will be measured by 

 the square of the fraction 



T_ tan(t-r) .^. 



tan(e + r)' - - - - - - \^) 



From these expressions it follows, that if a beam of light be 

 polarized in a plane making an azimuth A with the plane of in- 

 cidence, and if A' be the azimuth of. the polarization of the 

 reflected beam, 



tan A' = tan A — -7^ (. ... (3) 



cos [i—r) ^ ' 



If we now write Q= : i-o it is evident from (3) that the 



tan A' ^ ' 



value of Q obtained by measurement of A and A' shoiUd be con- 

 stant if FresnePs laws be true, and that it will not be constant 

 if the law be inexact. 



It is easy from (3) to deduce the following formulae, which 

 will serve to calculate the refractive index from measurements of 

 A and A' : 



Q-1 , . 

 tan7*= 7=z — rcot I 



/^= 



Q + 1 



sinz 



sinr 



(4) 



The following experiments were made with a rhomb of glass 

 made in Munich, the refractive index of which, corresponding to 

 a red very near the extreme red, was found to be 1*6229. The 

 experiments were made with white lamp-light from a Moderateur 

 lamp burning colza oil, and provided with a parabolic silvered 

 reflector; the polarizer consisted of a NicoFs prism, and the 

 analyser of a similar prism, made for me by M. Dubosc, and 

 without sensible deviation. 



Table I. 

 Incidence =39° 22'. 



The values of Q in this table are found from those of A and 

 A' in the first two columns, and the values of//, from equation (4) ; 

 both should be constant according to Fresnel's law, but it is 

 plain from the table that such is not the case. 



