ADVISORY COMMITTEE ON ASTRONOMY 1 35 



velocities of several hundred stars in the northern heavens, he is 

 now proceeding to South America to continue the work in the 

 southern sky. 



An important outcome of radial velocity determinations was the 

 discovery of the so-called ' ' spectroscopic binaries, ' ' which are double 

 stars whose component members are too close togethsr to be sepa- 

 rately distinguished with any telescope. The rapid orbital motion 

 of these bodies is betrayed by oscillations of the spectral lines. 

 When only one of the bodies emits sufficient light to produce a 

 spectrum, the only change observed is a to-and-fro motion of the 

 lines, which are displaced toward the red while the star is moving 

 away from us, toward the violet while the motion in the line of 

 sight is in the direction of the earth. If both components are lum- 

 inous, the lines of the compound spectrum are periodically doubled 

 through the effects of the orbital motion. Some fifty spectroscopic 

 binaries have already been discovered. So numerous are they that 

 Campbell estimates them to comprise one-seventh of all the stars. 



Researches on the radial velocity of stars with powerful instru- 

 ments are in progress at the lyick, Yerkes, Potsdam, Meudon, Cam- 

 bridge, Pulkowa, and Cape obsen^atories. At present most of the 

 work is confined to the brighter stars, but it has been found at the 

 Yerkes, Lick, and Potsdam observatories that the velocities of stars 

 as faint as the eighth or ninth magnitude can be measured with fair 

 precision with large telescopes. As the velocities of a great number 

 of stars must be ascertained for use in studies of the structure of the 

 universe, as well as in determinations of the motion of the solar sys- 

 tem in space, it is evident that a system of cooperation between ob- 

 servatories engaged in the work should soon be arranged. Hitherto 

 the necessity of developing methods and eliminating sources of sys- 

 tematic error has rendered independent work desirable, though cer- 

 tain standard stars are being sj'stematically observed through mutual 

 agreement of the above-mentioned obser\^atories. Within a short 

 time the period of preparation and experiment will have passed, 

 after which a general plan of cooperation should be adopted. 



Two important fields of stellar spectroscopy have yet to be noted. 

 The study of the spectra of variable stars, particularly those of long 

 period, for which important results have already been obtained by 

 Fleming, Belopolsky, Campbell, and others, promises an abundant 

 harvest of results, if sufficiently powerful telescopes can be used for 

 the purpose. In general, spectroscopic and photometric observa- 

 tions o£ these stars should be made simultaneously. The other 



