190 



ANNUAL REPORT SMITHSONIAN INSTITUTION, 1913. 



tioii of Prof. E. C. Pickering, with regard to the nature of their 

 spectrum. The principal groups of the Harvard classification are 

 desigTiated by the letters B, A, F, G, K, M. The peculiarities of 

 these types of spectra are indicated in the accompanying plate 3. 

 We see the progressive greater complexity of the spectra from type 

 to type. Campbell points out the very interesting fact that the more 

 complex the star spectrum, the greater the velocity of the star in 

 space, with regard to a point so fixed that the algebraic sum of the 

 velocities of all the stars with respect to it is 0. The same conclusion 

 is derived independently by Boss from a consideration, not of radial 

 motions, but of thwart motions of the stars. The results of Campbell 

 and Boss are compared in the following table. We assume for Boss's 

 results as for Campbell's that the velocity of the solar system toward 

 its apex of motion is 19.5 kilometers per second, thus the angular 

 motion observed by the telescope may be converted into its linear 

 equivalent. Unfortunately the grouping of stars by the two ob- 

 servers is different as regards the subclasses of the Harvard classifi- 

 cation. 



• For 132 of these stars the radial velocity is estimated. 



The reader must note that the results, both of Boss and of Camp- 

 bell, relate only to a certain component of the motion of the stars. 

 Ill the case of Boss it is derived from that component of the proper 

 motion, which is at right angles to the solar pathway; and in that 

 of Campbell it is that component of the radial motion which is in the 

 plane of the star and the solar pathway and is at right angles to the 

 solar motion. If it is assumed that the stars have no preference for 

 motion in one direction rather than another and that they are well 

 distributed over the whole celestial sphere it follows that the values 

 above given from both observers are but half the average velocity of 

 the group of stars, considering their motions in the real directions 

 which they have in space, and not merely the components of motion 

 found by Boss and Campbell. Thus we find for stars of group G, 



