1879.] some Experiments on Metallic Reflexion. 247 



In order to ascertain the probable error of the mean principal 

 incidence and azimuth as determined with Plate 1, the measure- 

 ments were repeated with Plate 6 ; the difference between the two 

 values of the principal incidence, according as one or other of the 

 neutral axes of the plate was in the plane of incidence, being least, 

 and therefore the retardation for red light differing least from 90° for 

 Plates 6 and 7. 



Quarter Undulation Plate at A. 



Plane of polarization Principal Principal 



of incident light. incidence. azimuth. 



+ 45 78 25 +35 ()' 



77 58 , 35 35 



77 42 35 05 



78 26 35 40 



-45 .... 78 24 -36 45 



77 47 . . 35 30 



77 47 36 15 



77 59 36 



Mean 78 03 35 43 



Quarter Undulation Plate at B. 



+ 45 73 57 -34 40 



73 17 36 10 



73 52 35 40 



73 35 35 20 



-45 74 30 +35 20 



73 05 ; . . 36 05 



73 37 34 35 



74 07 35 



Mean 73 45 35 21 



Similar measurements, which were about as concordant, were made 

 with the gold plate in water and carbon bisulphide. The numbers in 

 the table being the means of eight observations, four with the prin- 

 cipal section of the polarizing nicol inclined to the right, and four 

 with it to the left of the plane of incidence. 



f Plate at A. . 78 03 .... 35 43 



Gold in air <^ „ B. . 73 45 .... 35 21 



L Mean value. 75 54 35 32 



