EPIPOLIC DISPERSION OF LIGHT. 153 



As it has been clearly shown that a beam of white light from which certain 

 rays have been separated by epipolic dispersion is no longer susceptible of producing 

 the epipolic phenomena, it would seem a natural and almost a necessary conclu- 

 sion, that the rays so separated ought to be ivholly, or in a very high degree, so 

 dispersed when incident on an epipolizing surface. But the whole history of physical 

 optics is one continued warning against such seeming logical conclusions ; and in 

 this case also the conclusion is not borne out by fact. Thus in Exp. 2 and 4, abun- 

 dance of rays internally dispersed must of necessity have been incident on the new 

 surface presented to them, yet no fresh dispersion whatever took place. I may add 

 too that in experiments made with considerable care to exclude all other light from 

 incidence on a quiniferous surface, but such as had originated in epipolic dispersion, 

 I have not succeeded in obtaining any indication of their susceptibility of being a 

 second time so dispersed. Though from the obscurity of such rays as compared with 

 direct light, these trials can hardly be considered as proving a negative, yet they 

 certaiidy go very far towards proving the absence of any peculiar susceptibility in 

 those rays to this particular affection. 



J. F. W. Herschel. 

 Collingivoodi March 1,1845. 



Note added during the Printing. — Professor Graham has had the kindness to 

 transmit to me a specimen of an alkaloid, extracted from the brown coat of the seed 

 of the chestnut, to which the name Esculine has been given, which possesses in per- 

 fection the property of epipolic dispersion when in dilute solution, in which state it 

 precisely resembles quinine. The same eminent chemist refers also to a peculiar oil 

 called Colophene, formed by the regulated action of sulphuric acid on oil of tur- 

 pentine, which by his description of its phenomena, must also be an epipolizing liquid 

 of a similar character. 



J. F. W. H. 

 May 12, 1845. 



MDCCCXLV. 



