in the Neighbourhood of the Polarizing Angle, 



and the relation between S0 t and Bfi is 



sin 6 S/JL 



cos 6 l B0 l = 



^ 



or 80^—8/jb/p 2 , 



since cos 6 X = sin 0. Thus 



T/S= ^co S (^-^)- • • • • • (7) 

 In the case of water, 



/& G =1-341, /Lt B = l-331, S/i = '010. 



From these data, 



T 2 /S 2 = 3'46xl0- 5 ; 



showing that if the spot is central for Fraunhofer's line B, 

 the illumination at the centre for G is more than half as great 

 as is found (6) for B at the upper and lower limbs of the sun. 

 A considerable development of colour is thus to be expected, 

 when the band is well formed. 



The band may be achromatized with the aid of a suitable 

 prism, held between the nicol and the eye, but of course at 

 the expense of introducing colour at the upper and lower 

 limbs of the sun. I had at my disposal a glass prism of 10°. 

 This diminished, but could not annul, the colour when held 

 nearly in the position of minimum deviation ; but by sufficient 

 sloping the band was practically achromatized. When more 

 dispersive materials, e. g. benzole or bisulphide of carbon, 

 were substituted for water, the development of colour is very 

 great, and in the case of the latter made it impossible to judge 

 of the perfection of the band. The above-mentioned glass 

 prism was of course quite insufficient for compensation. 



The magnitude of these chromatic effects is given at once 

 by Brewster's law, which we may write in the form 



tan(0 + S0)=/*H-fy*. 

 Thus ^ 



S<?=oos»tfS /1 = r ^ ?l .... (8) 



which gives the angular displacement of the centre of the 

 dark band, due to the change from fi to /jl + S/jl. Let us 

 inquire what small angle (*") must be given to a prism of the 

 same material in order that it may be capable of compensating 

 the colour. The deviation D is equal to (/u — l)i, so that 

 .&D = 8/i.i. Hence, if SD=S0, 



