278 Mr KELLAND, ON THE MOTION OF 



I should, however, prefer retaining the two media above considered, 

 for reasons which I shall presently adduce. 



46. The conclusion at which we have arrived will be strengthened, 

 if we can shew that the value of Jr + C' is considerable, in cases in 

 which we have good reason to attribute to it effects, the magnitude 

 of which is determined by experiment. Such effects appear to me 

 to take place in the case of light. I must here refer to my INIemoir 

 on that subject, as I am not aware that it has elsewhere been pursued 

 into detail. 



By a reference to that Memoir, it appears that, representing the 

 square of the velocity by a series of terms of the form 



q I 



the quantity I is for all excepting two of Fraunhofer's substances 

 negative. 



This curious result is by no means attributable to the sum of the 

 series which represents it on the hypothesis of an ordinary vibration, 



^ = - 2/32 - sin^ — f 1 - ~] . The coefficients would thence 



(If r' 2 V r^ J 



be the same for all substances independently of a common factor. 



^'^- dt 



47. The explanation which I ventured to suggest of this anomaly 

 was one of which I have seen no reason since to repent. At the 

 same time I fear in repeating it that it may be pronounced incon- 

 sistent with facts. It is that the remaining terms in the expansion 

 must be retained: an hypothesis which seems to imply that the re- 

 fraction will depend (though probably very slightly indeed,) on the 

 inte7isHy of the light. 



This hypothesis is, I suppose, totally unsupported by any experi- 

 mental evidence, but it so readily solves the difficulty that I should 

 not be justified in leaving it unnoticed. 



Suppose then we adopt the hypothesis and examine the equations 

 of motion which thereby result. 



