1885.] Phenomenon of Crystalline Reflection. 177 



red through green to blue. Not unfrequently, however, the initial 

 tint may be green or blue, and on approaching a grazing incidence 

 we may get red or even yellow mixed with the blue, as if a second 

 order of colours were commencing. 



4. The colours are not in any way due to absorption ; the trans- 

 mitted light is strictly complementary to the reflected, and whatever 

 is missing in the reflected is found in the transmitted. As in the 

 case of Newton's rings, the reflected tints are much more vivid than 

 the transmitted, though, as will presently appear, for a very different 

 reason. 



5. As Dr. Herapath remarked to me long ago, the coloured light is 

 not polarised. It is produced indifferently whether the incident light 

 be common light or ligh t polarised in any plane, and is seen whether 

 the reflected light be viewed directly or through a Nicol's prism 

 turned in any way. The only difference appears to be that if the 

 incident light be polarised, or the reflected light analysed, so as to 

 furnish or retain light polarised perpendicularly to the plane of inci- 

 dence, the white Light reflected from the surface, which to a certain 

 extent masks the coloured light, is more or less got rid of. 



6. The character of the spectrum of the reflected light is most 

 remarkable, and was wholly unexpected. A direct- vision hand spec- 

 troscope was used in the observations, and the crystal was generally 

 examined in a direction roughly perpendicular to the plane of sym- 

 metry ; but it is shown well through a wide range of azimuth of the 

 plane of incidence. No two crystals, we may say, are alike as to the 

 spectrum which they show, but there are certain features common to 

 all. The remarkable feature is that there is a pretty narrow band, or 

 it may be a limited portion of the spectrum, but still in general of no 

 great extent, where the light suffers total or all but total reflection. 

 As the angle of incidence is increased, these bands move rapidly in 

 the direction of increasing refrangibility, at the same time increasing 

 in width. The character of the spectrum gradually changes as the 

 angle of incidence is increased ; for example, a single band may 

 divide into two or three bands. 



The bands are most sharply defined at a moderate angle of 

 incidence. When the angle of incidence is considerably increased, 

 the bands usually get somewhat vague, at least towards the edges. 



7. The commonest kind of spectrum, especially in crystals prepared 

 on a small scale, which will be mentioned presently, is one showing 

 only a single bright band ; and I will describe at greater length the 

 phenomena presented in this case. 



When the angle of incidence is very small, the light reflected from 

 the reflecting surfaces of the crystal shows only a continuous 

 spectrum. As the angle of incidence is increased, while it is still 

 quite moderate a very narrow bright band shows itself in some part 



