8 Sir D. Brewster's Observations on M. Rudberg's Memoir. 



generalization the anomalous and capricious phoenomena which 

 light exhibits in its passage through transparent bodies." 



In comparing the results obtained by M. Rudberg with 

 calcareous spar with those obtained by Malus, we have been 

 struck with a discrepancy so great that we cannot find any 

 explanation of it. 



Malus found the two indices for quartz to be 1-55817 and 

 1 '54843. Now if we compare these numbers with those of 

 M. Rudberg, we shall find that they both correspond to a ray 

 towards the extremity of the green space between the lines 

 E and F. Thus 



Extraor. Ray. Diff. Ordin. Ray. Diff. 



Rudberg... F 1-55894 1-54965 



Malus 1-55817 ;' 1-54843 



E 1-55631 ^^^ 1-54711 ^^^ 



Hence Malus's extraordinary index corresponds to a ray 

 whose distance from F is 77; while his ordinary index cor- 

 responds to a ray whose distance from F is liJ2. This dis- 

 crepancy is not at all to be wondered at, and shows us how 

 uncertain are all measures of refractive powers unless they are 

 referred to the fixed lines in the spectrum. 



The case is widely different, however, with calcareous spar, 

 as the following table shows. 



Ord. Index. Diff. Extr. Index. Diff. 



Rudberg... C 1-65452 B 1-48391 



Malus 1-65429 i 1-48330 " 



Rudberg... B 1-65308 



Here the results of Malus and Rudberg are widely divergent. 

 Malus used the middle green ray of the spectrum [Theorie 

 de la Double Refraction, Deux. Parte, § 42), or that which 

 is half way between E and F, and yet his mean ordinary 

 index corresponds with one of the red rays in Rudberg's 

 observations ! Assuming the same ray in both observations, 

 the mean ordinary index will be as follows: 



Mean Ord. Index. Diff, 



Rudberg 1-66585 -.-.ep. 



Malus 1-6.5429 ^^^^ 



In the case of the extraordinary ray the discrepancy is still 

 more surprising. Here the mean index of Malus corresponds to 

 a ray distant 61 from B, and existing beyond the end of' the 

 visible spectrumfoi-med by light ofordinai-y intensity. Assuming 



