STUDIES IN STELLAR STATISTICS. II. 101 
In the discussion of these values I exclude the star N:o 684 
(Square E,) on account of the abnormal large value of dr, : dt (—9.s 
Sir). We now obtain 
| 
Mean value of dr, /di= + 0.35 + 0.15 
» DT AD di 0.2 ns . 
Denoting the dispersions by «a, and a, we further obtain 
a, = 1.4 + 0.11 , 
LC ses US CO 
so that the dispersions in the velocities of the stars are of the same 
magnitude along the radius vector and perpendicular to it. 
It was to be expected that «, would have been greater than 
«c, and that approximately «, and a, (the dispersion of the velocities 
perpendicular to the plane of the Milky Way) would have had the 
same value. This was probable for theoretical reasons as well as on 
account of the direction of the vertex. I think that such a distribution 
of the velocities will result from the discussion of the velocities of the 
less hot stars, though the stars of the spectral type B, perhaps on 
account of their large masses and their slow motion, have not yet 
reached such a state of equilibrium. 
