5 68 SCIENCE PROGRESS 



spectrograph, and of the precautions which are taken in order to secure accurate 

 results. There follow in the third chapter the applications of these principles 

 to various problems of the solar system, such as the period of rotation of the sun, 

 and of Saturn's rings, and also to individual stars. The fourth chapter deals 

 with the various methods of determining the solar motion from statistical 

 considerations of the proper motions of stars, whilst in the following chapter 

 the same problem is discussed from the radial velocity determinations. This 

 method has considerable advantages as compared with the older one. In it 

 the actual radial velocities are used, whereas proper motions are determined 

 in arc, and for given linear motions they vary inversely as the (unknown) stellar 

 distances : moreover, radial velocities can be measured very accurately in a short 

 period of time, whereas the accurate determination of a proper motion requires 

 a series of observations extending over a long interval. The one disadvantage 

 of the method is that, at present, the velocities of the faint stars cannot be found, 

 but with increase in the power of the instruments used this difficulty will be 

 largely overcome. It is probable the velocity of the solar system in space as 

 determined by Campbell is the most accurate yet made. 



The sixth chapter contains several applications of the results obtained to the 

 stellar system. By eliminating the solar motion so as to leave the peculiar 

 motions of the stars, it is found that the number of positive velocities is con- 

 siderably larger than the number of negative, whereas the numbers should be 

 very nearly equal. The residual average velocity belongs almost entirely to 

 the stars of class B, and can be explained if there is an average increase in the 

 wave-lengths of all lines utilised of '07 A : such an increase could be caused by 

 a pressure of from twenty to thirty atmospheres ; and this may be the correct 

 explanation, because in these stars the absorption bands are of considerable 

 breadth, as if widened by pressure. Such a pressure effect appearing in so 

 unexpected a manner is of peculiar interest, and may throw some light on our 

 knowledge of class B stars. In this chapter also is a conclusive proof that stars 

 of early spectral types are travelling slower than those of later classes, as the 

 following table vividly shows : 



On the other hand, stellar velocities are not functions of the visual magnitudes. 

 There is no indication that the fainter stars are travelling more rapidly than the 

 brighter. Another important result is drawn by the author from statistical con- 

 siderations, and by comparison with results previously deduced from Proper Motion 

 data, viz. that the stars of various magnitudes are more thoroughly mixed in space 

 than had been previously supposed. These are but a few of the important 

 problems discussed in this fascinating chapter. 



The last two chapters are concerned respectively with visual and spectroscopic 

 binary stars, and with variable stars, and give a conspectus of our present know- 

 ledge on these subjects, which the study of radial velocities has helped to increase 

 in no small measure. 



The treatment of such a vast subject is necessarily incomplete ; but the author 

 has well succeeded in showing how rich a field of investigation is being opened 

 by the study of radial velocities, which, taken in conjunction with proper motion 



