270 ^ ON LIGHT. 



ought to appear as drawn out into a short, coloured 

 spectrum in a certain definite direction. Light requires 

 forty-two minutes to reach the earth from Jupiter at its 

 mean distance. Supposing the rays of one end of the 

 spectrum the violet, for instance to travel faster than 

 those at the other (the red), a satellite undergoing 

 eclipse by immersion in the shadow of the planet ought 

 to change colour before extinction, from white to red 

 the last-emitted red rays lagging behind the violet on 

 their journey to the earth ; while at its reappearance a 

 blue colour ought to be first perceptible. 



(53.) Among the stars are many which vary periodi- 

 cally in brightness, and some of them undergo complete 

 extinction. As light takes several years to travel from 

 the stars, the difference in the times of arrival for any 

 sensible difference of velocity would amount to many 

 days, and would be quite sufficient to tinge the disap- 

 pearing and reappearing star with the hues belonging to 

 opposite ends of the spectrum. No such thing, how- 

 ei^er, is observed. Most of them retain their whiteness; 

 and though some do assume a deep-red colour when un- 

 dergoing extinction, or when at their minimum of splen- 

 dour, it is not changed to blue at their reappearance, 01 

 on their commencing augmentation of briglitness. 



(54.) The reflexion and refraction of light are, as we 

 have stated, accounted for on this theory by supposing 

 the particles of all material bodies, besides the attractive 

 force of gravitation, to be endowed with other forces, 

 both attractive and repulsive the latter extending to a 

 greater distance than the former, so as to constitute an 



