ON THE NATURE OF LIGHT AND COLOUKS. 463 



assigned by Newton to a property which he attributes to the particles of 

 light, and which he supposes to direct them in the species of refraction 

 which they are to undergo: but on any hypothesis, until \vc discover the reason 

 why a part of the light is at first refracted in the usual manner, and another 

 part in the unusual manner, we have no right to expect that we should un- 

 derstand how thes^ dispositions are continued or modified, when the process is 

 repeated. 



In order to explain, in the system of emanation, the dispersion of the rays 

 of different colours by means of refraction, it is necessary to suppose that all 

 refractive medmnis have an elective attraction, acting mOre powerfully pn the 

 violet rays, in proportion to their mass, than on the red. But an elective at- 

 traction of this kind is a property foreign to mechanical pliilosophy, and when 

 we use the term in chemistry, we only confess our incapacity to assign a mechani- 

 cal cause for the effect, and refer to an analogy with other facts, of which the 

 intimate nature is perfectly unknown to us. It is not indeed very easy to give a 

 demonstrative theory of the dispersion of coloured light upon the supposition of 

 undulatory motion ; but we may derive a very satisfactory illustration from the 

 well known effects of waves of different breadths. The simple calculation of the 

 velocity of waves, propagated in a liquid perfectly elastic, or incompressible, 

 and free from friction, assigns to them all precisely the same velocity, what- 

 ever their breadth may be: the compressibility of the fluids actually existing 

 introduces, however, a necessity for a correction according to the breadth 

 of the wave, and it is very easy to observe, in a river or a pond of consider- 

 able depth, that the wider waves proceed much more rapidly than the nar- 

 rower. We may, therefore, consider the pure ethereal medium as analogous 

 to an infinitely elastic fluid, in which undulations of all kinds move wi-th 

 equal velocity, and material transparent substances, on the contrary, as 

 resembling those fluids, in which we see the large waves advance beyond the 

 smaller; and by supposing the red light to consist of larger or wider undu- 

 lations and the violet of smaller, we may sufficiently elucidate the greater 

 refrangibility of the red than of the violet light. 



It is not, however, merely on the ground of this analogy that we may be 

 induced to suppose the undulations constituting red light to be larger 

 than those of violet light: a very extensive class of phenomena leads us<still 



