Section III, 1921 [i4i] Tr.\xs. R.S.C. 



The Radial ]'elocities of jÇ4 Stars 

 By J. S. Plaskett, W. E. Harper, R. K. Youxg, H. H. Plaskett 



This work, wliich will appear as \'ol. II No. 1, Publication of the 

 Dominion Astrophysical Observatory, has been the main work of the 

 observatory since actual observing commenced in May, 1918. 



The stars were selected, in co-operation with the Mt. Wilson 

 Observ^atory, from those in Boss's Preliminary General Catalogue 

 north of the equator, which had not been previously observed for 

 radial velocity. The programme, as prepared for this observatory, 

 consisted of 770 stars in the alternate (even) minutes of right ascen- 

 sion. Of these 770 stars 50 around the eighth photographic magnitude 

 or fainter were postponed for lower dispersion, leaving 720 stars to 

 he obser\-ed. As 183 of these 720 stars proved to be spectroscopic 

 binaries or otherwise unsuitable, there were left 537 constant velocity 

 stars, of which 3,287 plates were obtained and measured, for the 

 main list. In addition a list of the "gamma" velocities of 22 spectro- 

 scopic binaries, whose orbits have been determined here from 544 

 plates, and a table of the estimated velocities of 35 binaries from 

 206 plates are given, making a total of 594 stars. 



The observations were all made with single prism dispersion, 

 34 A to the millimetre at H7 until Aug. 12, 1919, and 29 A after that 

 date when a larger angle prism was substituted. All the stars whose 

 spectral type is FO or later, provided the lines are sharp, were measured 

 on the spectro-comparator, and the earlier type spectra on micro- 

 meter engines. Inter-comparison of computed and measured values 

 of the standards on the spectro-comparator gives confidence that the 

 resulting ^'elocities, with this instrument, are free from appreciable 

 systematic error, which has been confirmed by the agreement of the 

 velocities of some stars obtained in common here and at other ob- 

 servatories. For the micrometer measures, the wave-lengths of the 

 lines used have been obtained from the best known values and, as 

 the same wave-lengths have been used throughout, any future change 

 in wave-length can readily be applied to the velocities. 



As the average number of spectra obtained and measured per 

 star is 6.1, and as the accordance in these well defined stellar spectra 

 is generally excellent, the mean velocities should be reliable. Some 

 measure of the accuracy is given by the probable errors as determined 

 in the usual way from the residuals, from the mean velocity, of the 

 individual plates. An idea of their general magnitude may best be 



