Stellar Motions

The first two lines give the motions of the streams relative to the sun, but greater cosmical importance must be attached to the motion of one stream relative to the other, which is given in the third line of the table. The principal difference between bright stars and faint stars lies in their distribution between the two streams. For the bright stars the numbers belonging to stream I. and stream II. are as 3:2, for the faint stars the numbers in the two streams are practically equal. It is for this reason that the declination of the solar apex increases progressively with the faintness of the stars considered; stream II. is receiving more and more weight in the "mean of the stars." Various attempts have been made to give a dynamical explana tion of star-streaming. It was pointed out by H. H. Turner that the comets of the solar system have a preponderatingly radial direction of motion. It might well happen that the orbits of the stars about the centre of the stellar system show the same tend ency; in that case, assuming that we are situated at some distance from the centre, the directions to and from the centre would be favoured directions of motion. The vertex or the antivertex would then indicate the direction in which the centre of gravity of the system lies. Others have favoured transverse instead of radial star-streaming, so that the centre of the system would lie 9ou from the vertex; this has the advantage that it is possible for an oblate system of stars to continue in a steady state with trans verse, but not with radial star-streaming, but on the other hand it is difficult to see how such a state of motion could originate. Observational evidence as to the direction of the centre of the galactic system does not accord well with either radial or trans verse star-streaming. On the whole, Kapteyn's original idea seems the best ; we have around us two groups of stars which have come together and are moving through one another.

Stars of type B are aloof from the two star-streams. Their motions (relative to the mean of the stars) are smaller than those of other types, and show little systematic tendency of any kind. A number of stars of other types are associated with them and form a third stream, known as stream 0. In most parts of the sky it is difficult to disentangle stream 0 from stream I, but in regions where the circumstances of projection are favour able the third stream always appears clearly.

Rotation of the Galaxy.

If the galactic system is analo gous to a spiral nebula it is presumably rotating in its own plane; in any case its flattened form would lead us to suspect rotation. Several attempts have been made to detect such a rotation from the proper motions of stars, but it is difficult to be sure that the results are not vitiated by the systematic errors of meridian cata logues. A determination by C. V. L. Charlier, in 1913, gave a general retrograde rotation of the stars in the galactic plane of 0•35" per century; his more recent result is 0.24". The rotation appears now to be confirmed by better evidence afforded by the radial velocities. If the rotation were like that of a rigid body the radial velocities, as viewed from any point, would be unaf fected; but it is more reasonable to suppose that the angular velocity diminishes outwards from the centre of the system, as with the planets in the solar system. Then, if we take a group of stars surrounding the sun, those between us and the centre will be gaining on us and those outside us will be lagging behind.

This shearing motion distorts a square into a parallelogram, that is to say one diagonal is lengthening and the other diagonal is shortening; so if we observe stars in two opposite regions of the sky, corresponding to the direction of one diagonal, these should be receding from us, and along the other diagonal, 9o° away, the stars should be approaching us, on the average. J. H. Oort has found that this effect is quite prominently shown in the ob served radial velocities, and that the direction of the galactic centre deduced from it accords well with the direction generally accepted (deduced from the distribution of globular clusters). Moreover, taking different classes of stars, the positions found for the centre are very accordant, and the effect increases, as it should do, proportionately to the mean distance of the class. Our survey covers only a relatively small region of the galactic sys tem, but we can at least say that in this region the stars have systematic motions which shear the distribution in the same way as orbital motion about the centre of the galaxy would do. It is possible, however, that other physical explanations of this be haviour might be given.

Individual Motions.

After eliminating all systematic mo tion there remains the individual motion of the star itself. An interesting field of study arises from the correlation between in dividual speed and other characteristics, such as mass, spectral type, brightness. The earliest result of this kind was found by W. W. Campbell and J. C. Kapteyn in 1910; viz., a progressive change of mean speed with spectral type. Campbell's results for bright stars were : Type B, 6.5 km. per sec. ; A, I z ; F and G, 15; M, 17 km. per second. The figures refer to the radial com ponent only, and must be doubled to give the mean speed in three dimensions. At the time it was thought that this sequence of types represented the order of evolution, so that the progression implied that the speed of the stars increases with their age ; with the recognition of giant and dwarf stars this interpretation has become inadmissible. Actually type M consists of two distinct classes, viz., very diffuse stars supposed to be in the earliest stage of condensation from a nebula, and very dense stars sup posed to be at the end of the evolutionary sequence; both classes have high speeds, the latter being particularly large. Generally speaking, faint stars have greater speeds than bright stars; owing to the mass-luminosity relation this means that stars of small mass have greater speeds than massive stars. It has been sug gested that there is a law of equipartion of energy so that the average value of my' is the same for all classes of stars; this might have a dynamical explanation, since the speeds of the molecules of a gas distribute themselves according to the same law. Up to a certain point observations agree with the equipar tion law, but there are noteworthy exceptions; the massive B stars have low average speed as the law requires, but the most massive stars of all, type 0, have rather large speeds. The present position is that our knowledge of the factors determining the mean speed of different kinds of stars is made up of fragmentary correlations, and has not yet led to any generally comprehensive theory.