202 CARNEGIE INSTITUTION OF WASHINGTON. 



about 80 stars with excellent trigonometric parallaxes and about 130 other 

 stars of well-known proper motion which have been observed recently. No 

 selection has been made except on the basis of apparent magnitude. Radial 

 velocities are known for many of these stars, so that derivations of mean 

 parallax are possible from pecuHar as well as parallactic motions. 



A preliminary study of the correlation between the intensity of the principal 

 line used in the spectroscopic study of these stars, the enhanced line of stron- 

 tium at X4215, and the reduced proper motion showed an excellent degree of 

 correspondence. On the other hand, the calcium line at X4455, the intensity 

 of which serves admirably to distinguish between giaijt and dwarf stars of this 

 type, is found to be comparatively insensitive to differences in absolute 

 magnitude among the giants themselves. This fact has been recognized 

 previously, and low weight has been assigned to it in this class of stars. Two 

 other hues, accordingly, have been selected of a character similar to X4215, the 

 enhanced strontium hne at X4077 and the enhanced iron hne at X4233, and 

 estimates have been made of their intensities in the spectra of the selected 

 list of stars. These intensities, divided into a number of groups, have then 

 been compared with the mean parallax as derived for each group from trigo- 

 nometric measures, parallactic motion, and peculiar motion. 



The results are conclusive in showing that the intensity of each of these lines 

 is correlated closely with absolute magnitude and serves as an excellent means 

 of deriving its value. They also show the close agreement for these groups of 

 stars of the mean parallaxes derived from trigonometric values on the one 

 hand and parallactic and pecuhar motion on the other. The accuracy of the 

 trigonometric parallaxes for these stars, many of which have small proper 

 motions, is thus shown to be quite comparable with that of the values derived 

 by the other two methods. The main conclusion to be drawn from the 

 investigation is that the dispersion in absolute magnitude among the K-type 

 stars is small and that the corrections to be appHed to the values pubHshed 

 previously are slight. 



Comparison of Spectroscopic and Dynamical Parallaxes. 



The hypothetical parallaxes of more than 1,600 visual double stars have 

 been calculated by Dr. Russell at Princeton University during recent years, 

 on the assumption that the mass of the system in each case is equal to that of 

 the sun. In order to correct this assumption and obtain dynamical parallaxes 

 based on the actual mean masses of stars of different spectral types, a com- 

 parison has been instituted by Russell, Adams, and Joy between the hypo- 

 thetical values for 327 of these stars and those derived spectroscopically at 

 Mount Wilson. 



The stars have been separated into giant and dwarf classes and divided 

 into groups according to spectral type. The interesting result then appears 

 that the relationship between the absolute magnitude Mi derived from the 

 hypothetical parallax and the ratio s/h of the spectroscopic and hypothetical 

 parallaxes is very nearly linear. All of the groups of stars agree in this 

 result, including the white dwarfs of types A and F. The residuals from the 

 formula 



sAi = 0.62 +0.045 Ml 



average about 6 per cent of both s and hi. A similar comparison between 

 s/hi and spectral type shows that in this case no simple relationship exists. 



