1879.] Double Refraction and Dispersion in Iceland Spar. 203 



peared it might be useful to have a series of results which would serve 

 in future as a test of any proposed theory of double refraction with- 

 out the necessity of a further recourse to experiment. It is therefore 

 largely with this object in view that I lay the results of my work 

 before the Society. 



I have sought to get a series of values of fi or — for the three hydro- 

 gen lines 0, F, g, in directions extending over an arc of 90° from the 

 optic axis. 



For this purpose I used four prisms, the edge of each was very 

 approximately perpendicular to the optic axis. 



The axis was inclined at angles of —32°, 14°, 38°, and 64° respec- 

 tively, to the normal to one of the faces of the prism, the angles being 

 considered positive when the optic axis and edge of the prism lie on the 

 same side of the normal. By this means I was enabled to obtain a 

 series of measurements in directions extending from about 5° on one 

 side of the axis to 100° on the other. Each of the measured quantities 

 used finally was the mean of two. The difference between the two was 

 in only 18 per cent, of the observations as great as 20", so that the 

 difference between the mean value taken and an extreme observation 

 was only rarely as great as 10". 



The spectrometer used was one by Grubb, of Dublin, kindly lent me 

 by Professor Stokes. It reads with verniers to 10". I found that 

 errors of 10" in the observed quantities taken so as to produce the 

 greatest effect in the result would produce at most variations of "00005 

 in the value of jx. 



Precautions were taken to maintain the room at as nearly as 

 possible a uniform temperature. 



The position of the faces of the prism with reference to the rhombic 

 faces of the crystal was determined by careful measurement. Each of 

 the results adopted in the calculations was the mean of ten. 



The results of the experiments are given in a series of tables. 



For the theoretical calculation we have, if yu. 1? be the principal 

 refractive indices, the angle between any wave normal and the 

 optic axis, 



1 _cos 2 ^_|_sin 2 

 ^ Pi 



We require then to find and yu 2 . 



lx Y is the refractive index for the ordinary ray, and was found from 

 observations on it in different positions of the prisms. yu 2 is the value 

 of fi for a wave normal at right angles to the axis, and was found 

 from observation on light travelling nearly in that direction. 



The results are given below. 



