282 IN STARRY REALMS. 



a velocity of a magnitude comparable with this that we 

 generally meet in the celestial regions. It may perhaps be 

 two or three times as great, or in some cases possibly as much 

 as ten times as great in any case, however, the velocities of 

 approach or recession are very small in comparison with 

 the velocity of light. Rarely, indeed, will the actual 

 velocity of a celestial body be so much as a thousandth 

 part of the velocity of light, and consequently no altera- 

 tions in the apparent quality of a ray arising from move- 

 ment can at the utmost do more than affect its wave length 

 by one-thousandth part of its total amount. 



It would be useful to obtain some concrete notion as to 

 the nature of the magnitudes with which we have to deal 

 when discussing the problem of celestial movements in the 

 line of sight. Let us look therefore at that ever-beautiful 

 object the solar spectrum, especially at the double line D, 

 which is caused by the presence of the element sodium. 

 The wave lengths of these lines expressed in ten-millionths 

 of a millimetre are, according to Professor Rowland, 

 5896-08, 5890-12 ; the difference, amounting as it does to 

 5*96, is about one- thousandth part of the whole amount. 

 We are thus able to form the general conception that the 

 greatest possible shift that could be caused in the spectrum 

 is not likely to exceed the length of the interval between 

 the two lines of sodium. It should be observed that in 

 this calculation I have taken an extreme value for the 

 velocity, and many of the velocities met with in space 

 will not be one-tenth as great. It therefore follows that 

 we must be prepared to make very exact measurements 

 when the phenomena with which we have to deal are of 

 such a delicate character that they do not extend to a 



