Polarization of Shy Light. 85 



"With ten or twenty plates the number of these portions is 

 very great and, though they are individuaUy faint, the aggre- 

 gate is considerable. Thus, of light polarized in the plane of 

 incidence about seven times as much actually traverses 

 twenty plates as is directly refracted through, when the angle 

 of incidence is forty degrees. But of light polarized in the 

 rectanguhir plane, that traverses the same plates, the re- 

 flexions only furnish about one seventh part. In the case of 

 ten plates with an angle of incidence of fifty degrees, the 

 corresponding proportions ?re three and a third times as 

 much and an infinitesimal fraction. Brewster used from six 

 to twelve plates and Rubenson twenty. 



Eubenson* avoided Brewster's error, but fell into another 

 connected with the reflexions. He determined not to rely on 

 theoretical calculations of the polarizing power of the plates, 

 especially as it would have been difficult to measure the 

 index of refraction with sufficient accuracy, and he employed 

 an excellent method, due to Arago, of standardizing his 

 instrument, i. e. measuring its depolarizing power for various 

 positions of the pile. He took a fairly thick plate of quartz, 

 cut parallel to the axis, and placed it with its principal plane 

 coincident with the plane of incidence on the pile. Behind 

 this he put a Nicol prism attached to the vernier of a graduated, 

 circle. Let the plane of polarization of the light incident on 

 the quartz make an angle with the principal plane ; then the 

 light emergent from the quartz is partially polarized, and the 

 ratio of the principal intensities is tan^(^. To secure com- 

 plete depolarization by the pile of plates, it is essential that 

 there should be no definite phase relationship between the 

 two components of the partially polarized light. If the 

 quartz plate had been thin this difficulty might have occurred; 

 but in a fairly thick plate the phase relationship is so com- 

 pletely changed by a slight alteration of wave-length that, 

 except with homogeneous light, the two components are 

 practically independent. 



So far, then, his method was admirable, but he never seems 

 to have remembered that the reflected light which comes 

 through the plates is more or less thrown to one side and 

 that, therefore, the aperture which admits light must be 

 large. If it be small only a proportion, and that a very 

 uncertain proportion, of the reflected light reaches the eve. 

 Rubenson had a large aperture in his observations on the sky 

 and a small one when he was standardizing. The Nicol 



* Memoire sur la Polarisation de la Lumiere Atmospherique, par Dr. 

 R. Rubenson. Upsal. 1864. To be had at Klemniing's Autiquariat, 

 Stockholm. 



