DEPARTMENT OF MERIDIAN ASTROMETRY. 



175 



in fair agreement with that indicated by the stars in general. Solutions 

 for preferential motion show that these stars follow closely the tendencies of 

 stars of other types in preferential motion towards Kapteyn's vertex, the 

 velocity figure being flattened toward the plane of the Milky Way. 



From the mean parallaxes derived by correlation of the proper-motion and 

 radial-velocity data, the following estimates of the mean absolute magnitudes 

 of the stars of the separate spectral classes were made: 



Spec. 



Misc. F, G, K 



Ma 



Mb 



Me 



Mabc 



Md 



M 



+0.4 

 +0.4 

 +0.3 

 +0.2 

 +0.3 

 ±0.0 



No. 



44 var. 

 11 var. 

 24 var. 

 26 var. 

 61 var. 

 89 var. 



Spec. 



Pec 



S.. 

 R. 



Na, 

 Nb 

 Np 



M 



+0.3 

 +0.1 

 -1.6 

 -1.4 



-2.4 

 -4.2 



No. 



6 var. 

 14 

 15 

 16 

 52 



9 



The tabulation indicates that, at maximum certainly, the long-period 

 variables must be classified as giant stars; that in those classes composed 

 wholly of stars whose light is subject to large variations there is a very small 

 range of mean absolute magnitude; and that the stars of Classes R and N, 

 composed of stars whose light is either apparently constant or subject to 

 comparatively small variations, are definitely more luminous, a conclusion 

 in harmony with the spectrographic evidence that these stars form a branch 

 of the curve of stellar evolution independent of that of the G, K, and M stars. 

 These results, in conjunction with those derived elsewhere for the non- 

 variable stars, indicate that there is no great range in mean luminosity for the 

 giant stars of types ranging all the way from F5 to S. If from this we are 

 permitted to reason that there can not be a great range in luminosity among 

 the individual stars within each class, the apparent magnitudes may be taken 

 as measures of distance. Comparisons of the parallaxes computed on this 

 assumption with available trigonometric parallaxes indicate that this 

 assumption is not unreasonable for the stars of the M classes. 



VELOCITY DISTRIBUTION. 



A part of the preparation of a paper. On the Real Motions of the Stars,^ by 

 Benjamin Boss, Harry Raymond, and Ralph E. Wilson, falls within this 

 year's report. Aside from the conclusions noted in previous reports, atten- 

 tion might be called to the expression of the unequal distribution of velocity 

 directions of high-velocity stars as a skewness of the velocity distribution per- 

 vading the motions of all classes of stars. The axis of greatest skewness lies 

 approximately in the Galactic plane in longitude 45° to 225°, the large motions 

 avoiding the former direction and preferring the latter. 



An attempt was made by Mr. Raymond to separate the skew distribution, 

 as projected upon the Galactic plane, into two Gaussian distributions, either 

 circular (two-stream) or elliptical (double-ellipsoidal), using Akesson's 

 method.- The distribution did not satisfy Akesson's criterion, and therefore 

 its skewness can not be explained by such a mixture. If actually due to a 



' Astronomical Journal, No. 820. 



*0n the dissection of correlation surfaces, Lund Meddelande, No. 73. 



