152 SIR J. F. W. HERSCHEL ON THE 



line to nearly half an inch from the surface, thus leaving it doubtful whether some 

 small amount of dispersion may not be effected in the interior of the medium at ap- 

 preciable depths. 



The narrow blue line above described was viewed through a Fraunhofer flint 

 prism. The spectrum was deficient at the red end by the totality of the purer and 

 less refrangible red, nearly the whole orange, and all the yellow. A rich and broad 

 band of fine green ligiit slightly fringed with red on the less refrangible side, passed 

 suddenly, on the more refrangible, to a copious indigo and violet without any inter- 

 mediately graduating blue. Either from want of sufficient brightness, or from some 

 other cause, no black lines were seen ; as far as mere illumination went, the spectrum 

 developed appeared continuous. 



It appears from this that no one prismatic ray in particular is selected for epi- 

 polic dispersion, but that a certain small per-centage of rays extending over a great 

 range of refrangibility are subject to be so affected, the less refrangible extreme being 

 however wholly excluded, as well as the majority of all below a mean refrangibility. 



The epipolic colour is more intense the more oblique the visual ray is to the 

 dispersing surface. This, which would be inexplicable on the supposition of the 

 dispersion being effected rigorously at the geometrical surface of the medium, is a 

 necessary consequence of its taking place within a superficial stratum of very small, 

 but appreciable thickness, or according to a law of intensity decreasing with great 

 rapidity as the depth within the medium increases. It has been already shown that 

 the dispersion is not confined to the interior of the liquid, but that a large portion of 

 the dispersed light is directed outwards, Exp. 4. The more oblique portions of this 

 (which are also the more intense) require, as is there shown, peculiar management to 

 render them visible. Those whose inclination to the dispersive surface is greater, 

 may also be subjected to ocular inspection, by carefully destroying all regularly re- 

 flected or accidental light. Thus, if on a surface of black paper two blots be made, 

 the one of water, the other of a solution of quinine, and if these be laid before a 

 window and viewed through a blackened tube in any direction but that of regular 

 reflexion, the water will appear perfectly black, the quinine feebly blue. But however 

 oblique to the surface the visual ray may be in this case, no great accession of inten- 

 sity takes place in the epipolic tint, for this obvious reason, that the dispersing 

 stratum being ivithin the medium, no ray dispersed by it can penetrate the surface, 

 which has not an inclination thereto exceeding 41° 22', at which angle, therefore, it 

 must cut the stratum, and cannot therefore traverse any great extent of it bodily. 



Hence, moreover, on the other hand, the internally dispersed light, at great obli- 

 quities to the surface (supposed in contact with air), will be reinforced by all that 

 portion which would have penetrated the surface and gone into the air but for the 

 law of total reflexion ; all the dispersed rays, that is to say, whose inclination to the 

 surface is less than 41° 22'. This consideration helps to explain the great comparative 

 intensity which the dispersed beam possesses under such circumstances. 



