248 



Sir J. Conroy on 



[Jan. 9, 





r Plate at A. . 



74 46 . . 



. . 36 





< „ B. . 



70 25 



. . 36 49 





1 Mean value. 



72 35 



. . 36 24 





r Plate at A. . 



71 37 



. . 36 59 



Gold in carbon bisulphide 



I » B:. 



68 26 



. . 36 43 





L Mean value . 



70 01 



. . 36 51 



The mean values of the principal incidence and principal azimuth 

 obtained with the two quarter undulation plates being different, it 

 was assumed that the errors of the means are as the squares of the 

 small errors of the plates, and that the errors of the incidences in 

 either position of the plate, and therefore the algebraical differences 

 or numerical sums of the errors in the two positions, that is, the 

 differences of the apparent principal incidence in the two positions, 

 as the first powers ; and therefore that the errors of the means are 

 as the squares of the difference of incidences in the two positions. 



Gold in Air. 



Plate 1. Plate 6. Correction. 



o / o / / 



Principal incidence 75 29 .. 75 54 . . +25 



azimuth 35 54 . . 35 32 . . -22 



Difference of principal inci- Q / , t t 



dence in two positions ... 9 28 or 568, 4 18 or 258. 



Thus the residual corrections to the res alts got with Plate 6 will 

 be to the difference on the results got by Plate 1 and Plate 6, as 258 2 

 to 568 2 - 258 2 , or as No. log 1-41491 to 1; this gives + 6' and -5' 

 making the corrected principal incidence and principal azimuth 76° 

 and 35° 27'. 



In a similar manner the means of the results got with the gold 

 plate in water and carbon bisulphide were corrected, the final results 

 being with red light. 



Principal incidence. Principal azimuth. 



Gold in air 76 .... 35 27 



water 72 46 36 23 



„ carbon bisulphide. ... 70 03 .... 36 48 



In order to determine the principal incidence and azimuth for gold 

 by an independent method, the one originally used by Sir David 

 Brewster was adopted ; the quarter undulation plate was removed, and 

 a second gold plate attached to the vertical stage in such a manner 

 that, whilst the plates remained parallel to each other, the distance 

 between them could be altered. The plates were so adjusted that 



