112 FKAGMENTS OF SCIENCE. 



or by the same absolute quantity. They must be re- 

 duced proportionately, instead of equally. If by the 

 act of reflection the waves of red light are split into 

 exact halves, then, to preserve the light white, the 

 waves of yellow, orange, green, and blue, must also be 

 split into exact halves. In short, the reduction must 

 take place, not by absolutely equal quantities, but by 

 equal fractional parts. In white light the preponder- 

 ance, as regards energy, of the larger over the smaller 

 waves must always be immense. Were the case other- 

 wise, the visual correlative, Hue, of the smaller waves 

 would have the upper hand in our sensations. 



Not only are the waves of ether reflected by clouds, 

 by solids, and by liquids, but when they pass from light 

 air to dense, or from dense air to light, a portion of the 

 wave-motion is always reflected. Now our atmos- 

 phere changes continually in density from top to bot- 

 tom. It will help our conceptions if we regard it as 

 made up of a series of thin concentric layers, or shells 

 of air, each shell being of the same density throughout, 

 a small and sudden change of density occurring in 

 passing from shell to shell. Light would be reflected 

 at the limiting surfaces of all these shells, and their ac- 

 tion would be practically the same as that of the real 

 atmosphere. And now I would ask your imagination 

 to picture this act of reflection. What must become 

 of the reflected light? The atmospheric layers turn 

 their convex surfaces towards the sun; they are so 

 many convex mirrors of feeble power; and you will 

 immediately perceive that the light regularly reflected 

 from these surfaces cannot reach the earth at all, but is 

 dispersed in space. Light thus reflected cannot, there- 

 fore, be the light of the sky. 



But, though the sun's light is not reflected in this 

 fashion from the aerial layers to the earth, there is 



