MOUNT WILSON SOLAR OBSERVATORY. 267 



Spectral Criteria of Absolute Magnitude. 



In the course of the comparison of the spectra of the large and the 

 small proper-motion stars, two differences of a systematic character 

 between these two kinds of stars have been observed in the line spectrum: 



1. The hydrogen lines are abnormally strong in a considerable 

 number of the small proper- motion stars. The effect is particularly 

 striking in the case of several stars which show strongly the titanium 

 bands characteristic of type M. The hydrogen lines in these stars 

 normally should be comparatively weak, but actually are as strong as 

 in type G5 of the Harvard classification. In many other stars showing 

 these bands the hydrogen lines are as strong as in normal K-type stars. 

 Although by no means certain as yet, it seems probable that abnormal 

 strength of the hydrogen lines will prove to be a characteristic feature 

 of all stars of great absolute luminosity. Measures of the radial 

 velocities given by the hydrogen lines are in sufficiently close agree- 

 ment with those given by other lines to indicate that the excess absorp- 

 tion is mainly, if not wholly, due to the stellar atmospheres, and not 

 to hydrogen gas in space. 



2. Certain spectrum lines are weak in the large proper-motion stars 

 and strong in the small proper-motion stars, and conversely. The 

 most notable cases of lines strong in the large proper-motion or low- 

 luminosity stars are four lines which are much strengthened in the 

 spectra of sun-spots, and which probably are low-temperature lines. 

 The most important case of a line strong in the high luminosity stars 

 is the strontium fine at X 4216, which is exceedingly prominent in the 

 spectrum of the chromosphere. The possibility of applying these 

 results to the determination of absolute magnitudes has been investi- 

 gated in the following way: 



The spectral types of 162 stars were first determined accurately from 

 a comparison of the intensity of the hydrogen lines with certain other 

 lines in the spectrum, the differences of intensity being estimated on 

 a definite scale, and this scale then reduced to the Harvard system by 

 means of stars of known type. The 162 stars were divided into two 

 groups — types Fg to Ge, containing 71 stars, and t^^pes Ge to Kg, con- 

 taining 91 stars. 



The absolute magnitudes of all of the stars were then computed, 

 from the parallaxes where they were available and from the proper 

 motions where they were lacking or very small. The means of the 

 absolute magnitudes were then formed separately for the low and the 

 high luminosity stars, and against these were plotted the differences 

 of intensity of those lines which showed the greatest variation with 

 absolute luminosity. Curves were drawn through the results and then 

 each curve was applied to the case of the individual star. In other 

 words, the absolute magnitude of the star was determined from the 

 relative intensities of the selected-spectrum lines. 



