Average Deduced 



Type of Number Radial Motion Correction 



Spectrum. of Stars. Velocity. of Sun. K. 



B 225 6.5 20.2 1-4-1 



A 177 II. o 16.8 -j-i.o 



F 185 14.4 -15-8 +0.1 



G 128 15.0 16.0 0.2 



K 382 16.8 -21.2 +2.8 



M 73 I7-I -22,6 +3.9 



The first and third columns chiefly concern us. The first shows the type of spectrum 

 on the Harvard classification, and in the order assigned for age or physical development 

 of the star: the third gives the average velocity in km. per sec. The fourth column shows 

 the velocity of our own sun deduced from the different groups, and the fifth a curious 

 general correction which seems to be required by the means of the groups. (If real it 

 would mean that the B stars for instance are moving generally outwards from the sun 

 at 4 km. sec.: but Campbell is inclined to interpret the column K as systematic error in 

 the measurement of the various types of spectrum.) 



The meaning of this striking association of movement with age, or at least with 

 spectral type, is not yet clear. In the astonishment immediately following the dis- 

 covery some rather wild suggestions were made, as for instance that a star was at birth 

 not subject to gravity, but only gradually came under its influence. It seems probable 

 however that some simpler explanation than this can be found, and one such is indicated 

 below: but the chief requirement at present is further investigation. If the B stars 

 alone are considered, especially those with large velocities, H. Ludendorff has shown 

 that their velocities follow J. N. Lockyer's classification of probable life-history in a 

 striking manner, those on the ascending branch of Lockyer's curve having positive 

 velocities (outwards from our sun) and those on the descending branch negative (in- 

 wards). There may be other facts of this kind still to come. 



It is a further outcome of the Lick observations that the stars of types A and B tend 

 to move parallel to the plane of the Galaxy. Assuming that this is exactly true, H. C. 

 Plummer has shown that the distances of the stars and their complete movements can 

 be deduced from a combination of telescopic and spectroscopic observations. The 

 deductions will be exactly correct only if the assumption is exact, but if it is roughly 

 true the deductions will also be roughly true. Plummer has accordingly deduced the 

 complete motions of more than 1 50 stars with the result of finding them fall into several 

 groups of stars with similar motions, the association being in some. cases strikingly de- 

 fined. It had already been suggested by J. Halm, in discussing Kapteyn's hypothesis 

 of two star streams (for which see article " Star " in the E. B.), that there must be a third 

 stream present moving towards a vertex nearly coinciding with that of one of the former 

 two: Plummer 's results suggested a still greater number of streams towards that vertex, 

 though they fail to indicate the opposite vertex. Evidence of the complexity of stellar 

 movements has also come from other quarters. L. Boss's discovery of a special cluster 

 in Taurus, and Ludendorff 's confirmation of the Ursa Major clusters, were followed by 

 the detection, by three investigations simultaneously, of two groups of stars of the. helium 

 type, having nearly equal and parallel motions. One of these is spread over a large 

 area in Scorpio and Centaurus: the other consists of about seventeen stars in and near 

 Perseus. Finally it has been noticed that the helium stars generally do not conform to 

 the two star-streams. 



Amid this increasing complexity there is one suggestion of a unifying principle. It 

 has been pointed out by H. H. Turner that if the stars around us form a limited system 

 (as has been suggested by others) isolated by distances large compared with its own 

 dimensions from other such systems, then the gravitational attraction of the system as 

 a whole will result in a central attraction on the individual members. Further, that if 

 the system started from rest, the movements of individual stars would approximate to 

 oscillations backwards and forwards through the centre. Putting aside for a moment 

 the obvious difficulties of congestion at the centre, then there would at any given mo- 



