THE MOVEMENTS OF THE STARS. 285 



Now, though that atmosphere may be sensibly transparent, 

 it nevertheless levies a certain toll on the light passing 

 through it ; in fact, it stops some of the light. 



Let us suppose, for the sake of argument, that the atmo- 

 sphere surrounding the star was solely composed of hydro- 

 gen gas. This gas would be almost perfectly transparent, so 

 tar as most of the rays are concerned. It is essential, in 

 this point, to pay attention to the wave lengths ; most 

 rays will be permitted to fly through unimpeded, but the 

 passage of certain particular rays will be opposed, if not 

 altogether prohibited, by the action of the hydrogen. 

 Take, for instance, the ray whose wave-length is 4860*72, 

 that particular ray will be more or less stopped ; if its wave 

 length had been a few units greater, or were a few units less, 

 it would have been allowed to pass, but being what j't is, its 

 passport is refused. The same may be said of certain 

 other wave lengths, The wave length of 6562*1 units, 

 also will be stopped ; while most rays more refrangible or 

 most rays less refrangible are allowed to make their out- 

 ward journey without hindrance. Here is indeed a 

 remarkable property possessed by a so-called transparent 

 gas ; it picks out here and there certain wave lengths, and 

 refuses all passage to rays which radiate in these for- 

 bidden periods, while rays of all other wave lengths pursue 

 their way uninterfered with. At first it might seem that 

 the rays which were stopped were selected apparently by 

 <japrice. Hydrogen, for instance, of which we are now 

 particularly speaking, stops certain rays in all parts of the 

 spectrum. But now comes the important discovery to 

 which we have referred the discovery which in fact 

 renders spectrum analysis so important to the astronomer. 



