140 CORRELATION OF PHYSICAL FORCES. 



reach the earth ; so that we see them as they existed ten 

 years ago. The distance of most visible stars is probably far 

 greater than this, and yet their brilliance is great, and in- 

 creases when their rays are collected by the telescope in pro- 

 portion ceteris paribus to the area of the object-glass or spec- 

 ulum. There is, however, an argument of a somewhat spec- 

 ulative character, by which light would seem to be lost or 

 transformed into some other force in the interplanetary spaces. 



Every increase of space-penetrating power in the tele- 

 scope gives us a new field of visible stars. If this expansion 

 of the stellar universe go on indefinitely and no light be lost, 

 then, assuming the fixed stars to be of an average equal 

 brightness with our sun, and no light lost other than by diver- 

 gence, the night ought to be equally luminous with the day ; 

 for though the light from each point diminishes in intensity 

 as the square of the distance, the number of luminous points 

 would fill up the whole space around us ; and if every point 

 of space is occupied by an equally brilliant point of light, the 

 distance of the points becomes immaterial. The loss of light 

 intercepted by stellar bodies would make no difference in the 

 total quantity of light, for each of these would yield from its 

 own self-luminosity at least as much light as it intercepted. 

 Light may, however, be intercepted by opaque bodies, such 

 as planets ; but, making every allowance for these, it is diffi- 

 cult to understand why we get so little light at night from the 

 stellar universe, without assuming that some light is lost in its 

 progress through space not lost absolutely, for that would be 

 an annihilation of force but converted into some other mode 

 of motion. 



It may be objected that this hypothesis assumes the stel- 

 lar universe to be illimitable : if pushed to its extreme so as 

 to make the light of night equal that of day, provided no 

 stellar light be lost, it does make this assumption ; but even 

 this is a far more rational assumption to make than that the 

 stellar universe is limited. Our experience gives no indica- 



