EPIPOLIC DISPERSION OF LIGHT. 149 



from which its former contents were in no way distinguishable by an eye situated 

 behind it. Being then replaced as before, so as to intercept the light incident on 

 the test cylinder, the epipolic colour was produced, exactly as if nothing had been 

 interposed ; a trifling difference of intensity only excepted, which arose from the 

 glass used not being wholly devoid of colour. 



Exp. 4. This experiment was varied so as to present a result disengaged from this 

 slight source of uncertainty, and perfectly decisive. A cylindrical jar was coated exter- 

 nally with black paper round three-fourths of its circumference, as was also its bottom, 

 and a ring of the same paper was carried round the cylinder at the bottom so as to cut 

 off light from being internally reflected on its base. In it was set upright a test cylinder 

 of the solution, and the jar was then filled with pure water rising considerably above 

 the solution in the cylinder. When exposed to light as usual, the epipolic tint was 

 finely seen. But on emptying out the water, and introducing in its stead an equal 

 quantity of the quiniferous solution, the tint in question was completely destroyed, 

 whether the surface of the cylinder was viewed from within or from without, proving 

 evidently that no rays susceptible of epipolic dispersion had reached its surface. 

 This result was rendered the more remarkable by an effect of contrast. The external, 

 upper portion of the cylinder, above its liquid contents, but below the level of the 

 liquid in the jar, reflected to the eye (or rather the air within it reflected) a pretty 

 strong blue gleam, being no other than the epipolically dispersed light of the anterior 

 surface of the liquid in the jar ; while all below (being glass in contact with the 

 liquid on both sides and so deprived of reflective power on both surfaces) was com- 

 pletely dark and almost invisible. 



When the interior test cylinder was sloped backwards from the incident light at 

 an angle of about 70° to the horizon, a beautiful and instructive feature was deve- 

 loped. In this situation of things, the interior liquid being as usual the quinine solu- 

 tion, and the exterior pure water ; to an eye perpendicularly over the surface, the whole 

 anterior portion of the cylinder from below upwards to the surface of the interior liquid, 

 appeared coated as it were internally with a most delicate and beautiful blue film of 

 extreme tenuity and perfect transparency, presenting a singular ghost-like appear- 

 ance, easier produced than described. This being seen through the cylinder, by an 

 eye situated externally to its prolongation, affords a proof that the epipolic dispersion 

 takes place in all directions : but except in this mode of viewing it the rays dispersed 

 outwards cannot reach the eye, or not in abundance (for which a very oblique inci- 

 dence is required), being at such an incidence internally and totally reflected by the 

 outer surface of the glass. To see this to advantage an eye-tube internally blackened 

 should be used to guard the eye from extraneous light. Such a tube indeed is ge- 

 nerally advantageous in all these experiments. 



If, instead of water, the test cylinder be plunged into a solution of quinine, all 

 else remaining as before, the blue film in question totally disappears. I tried a great 

 many other liquids, all in fact which I had at hand in sufficient quantity and colour- 



