242 FRESNEL. 



will manifest itself by an increasing deviation. Refrac 

 tion thus becomes a sure means of comparing the veloci 

 ties of all sorts of rays. In following out this research 

 with experimental means so precise that they would 

 mark differences of one fifty-thousandth part of the whole 

 amount in question, we have been able to ascertain that 

 the light of all the heavenly bodies, of our fires, of 

 tapers, and lamps, (with double currents of air,) and 

 even more, the feeble rays emitted by glow-worms, all 

 go through 77,000 leagues in a second, as well as the 

 burning light of the sun. 



It is easy to conceive in what way this result is a 

 mathematical consequence of the system of waves, if we 

 only remark that all the notes of the musical scale are 

 propagated with equal rapidity through the air, whether 

 they originate from the voice of a singer, from the metal 

 lic string of a piano, the catgut of a violin, the glass sur 

 face of an harmonica, or the metallic sides of the great 

 pipe of an organ. Now, there is no reason why the 

 luminous notes (if I may be pardoned the expression) 

 should proceed differently in asther. On the hypothesis 

 of emission the explanation is not so simple. II light is 

 composed of material elements, it would find itself sub- 



The undulatory view of refraction depends entirely on the assump 

 tion that the velocity must be diminished in the denser medium; but 

 as the refraction is proved to be constant for all velocities, this dimin 

 ished velocity must be always in a constant ratio to the original veloc 

 ity. This is the condition to which our conception of sether must 

 conform. As to the fact of a rtlardation, that has been directly 

 proved by another beautiful experiment devised by Arago, but car 

 ried out by the experimental skill of Foucault, on the principle of 

 Wheatstone s i evolving mirror, which, if it received one of two rays 

 at the smallest interval of time after another, would reflect it in a dif 

 ferent direction. Thus the existence of any retardation in one of the 

 rays would be manifested. 



