DEPARTMENT OF MERIDIAN ASTROMETRY. 235 



D = +35° from proper-motions; A = 266°, D = +25° from radial veloci- 

 ties; or for general purposes, the position A = 270°, D = +30°. 



Aside from the results published, solutions have also been made for 

 subdivisions of each type according to magnitude. The completed 

 solutions for the two subdivisions 0'.' to 5y2 and 5''.'3 to 6'|0 are ready 

 for publication, but for stars of fainter magnitude the types are not 

 available. Publication will therefore be delayed until the completion 

 of the Harvard classification. 



Directly related to the foregoing investigation is one by Mr. H. 

 Raymond, dealing with preferential motion according to type. Mr. 

 Raymond based his investigation on 6,033 proper-motions from the Pre- 

 liminary General Catalogue, treating them according to Schwarzschild's 

 method. The proper-motions were successively grouped according to 

 type, according to hem.ispheres, and again by type according to galactic 

 and non-galactic latitudes. A summary of his results follows: 



1. The vertices for the various groups lie in a band making a small 

 but well-marked angle with the galaxy, practically in the direction of 

 solar motion. The vertices for the early types A and F and those for 

 the late types K and M lie somewhat separated from each other near 

 the two ends of this band. Type G, as in the case of solar motion, 

 is anomalous. As a similar phenomenon is exhibited in the determi- 

 nations of the apex of solar motion, the question arises whether the 

 determination of apex and vertex are completely independent of each 

 other. Mathematically the recognized methods seem to be sound; 

 but they rest upon assumptions in regard to the areal and radial dis- 

 tribution of the stars, the similarity of motions in different regions, 

 and the laws of distribution of those motions, which can hardly be 

 true, save in the roughest way. The distribution of vertices is nearly 

 along a meridian. If, however, it were due to errors in the proper- 

 motions in declination, it should also appear in the solution by hemi- 

 spheres. It is difficult to believe that errors of sufficient size could 

 exist, varying with type but not with position in the sky. 



2. The range of galactic and non-galactic solutions, for each group, 

 falls, when at all large, into the same belt. Type G is again anomalous, 

 in regard to both the amount and the direction of the range. Anomalies 

 seem more frequent for non-galactic than for galactic stars. These 

 results can claim no great accuracy, especially for the smaller types. 



3. The ratios of the mean velocity in the direction of preferential 

 motion to that at right angles to that motion show, rather uncertainly, 

 that the preferential element of motion is stronger in the earlier types, 

 and within the earlier types, for non-galactic stars; and decisively that 

 it is stronger for large proper-motions. 



4. Attention has been called to the fact that the proper-motions of 

 the earlier type stars are greater in the non-galactic than in the galactic 

 regions. It has been suggested that the velocity-figure for these stars 

 might be an ellipsoid of three unequal axes, thin in the direction of the 



