STUDY OF ORES AND METALS. 441 



since the apparatus is more complicated and the attainable accuracy 

 is not increased, such devices are not recommended for practical 

 work. 



Effect of Reflecting Prism or Plate on the Character of Light. — 

 An extended series of experiments with different kinds of reflectors, 

 total reflecting prisms, silver surfaces, silvered glass mirrors with 

 central aperture or a small disk mirror in center of field, or an Abbe 

 reflecting cube, was carried out to ascertain if possible which type 'is 

 the best for such work, especially for use with the second group of 

 methods. The general equations state that in case the azimuth of 

 the plane of polarization of the incident wave is zero or 90°, the 

 reflected beam is still plane polarized ; practical experience with such 

 reflectors states, however, that the reflected beam always shows traces 

 of elliptical polarization even when the plane of polarization is hori- 

 zontal or vertical. This is no doubt due in the case of reflecting 

 glass surfaces to internal reflections ; with a blank metal reflecting 

 surface, such as the fresh silver side of a silvered mirror, it may re- 

 sult from strains in the outer film. Be the cause what it may, in no 

 experiment was the elliptic polarization entirely removed, and the 

 accuracy of the settings was correspondingly diminished. Expe- 

 rience with the different types of reflectors did not demonstrate 

 marked superiority of any one particular type. It is, of course, es- 

 sential that the plane of polarization of the incident beam be strictly 

 horizontal or vertical in order to reduce to a minimum the amount of 

 elliptic polarization present. 



Measurement of Phase Differences in Reflected Waves. 



The curves of Fig. 6 prove that in general the phase difference 

 between the components of the reflected waves normal and parallel 

 to the plane of incidence is of a low order of magnitude, so low in 

 fact that it may in general be neglected. The actual measurement 

 of the phase difference is best accomplished by the standard method 

 with a Babinet compensator. In view of the slight differences to be 

 observed, however, this method is not to be recommended for prac- 

 tical diasfnosis. 



