THE MOVEMENTS OF THE STARS. 281 



canvas as that which suffices for our mere visual impres- 

 sions of colour. When we study the refinements of spec- 

 troscopic analysis, the indications of celestial movement 

 become visible enough, even though they are insufficient 

 to turn a green into a blue, or a yellow into a red. The 

 development of this subject is one of immense interest, 

 but it also possesses some little difficulty ; and I therefore 

 entreat your most particular attention. 



I shall revert again to the supposition of a green star, che 

 radiation from which is monochromatic. If the light from 

 such an object were examined through a spectroscope the 

 appearance presented would not be that of a band of varied 

 colours which form the glories of the ordinary spectrum ; it 

 would be rather a single line of vivid green light, and occu- 

 pying a particular position in the spectrum as defined by 

 its wave length. If the body were moving towards us the 

 line would shift towards the blue end. If the body were 

 moving from us the lines would shift towards the red end. 

 Let us suppose that our spectroscope is furnished with a 

 measuring apparatus by which we are enabled to deter- 

 mine the amount of the shift with the necessary accuracy. 

 Suppose, in fact, that the wave lengths of the light, when 

 the object is relatively at rest, be the 30,000th of an inch, 

 but that when the object is moving towards us it is found 

 that the vibrations are 30,003 in the inch; it then follows 

 that the fact of this body's movement towards us has caused 

 us to receive one ten -thousandth part more vibrations per 

 second than when the body was at rest. It therefore fol- 

 lows that the velocity with which the body is dashing to- 

 wards us must be a ten- thousandth part of the velocity of 

 light in other words, about eighteen miles a second. It is 



