ROTATION OF THE GALAXY EDDINGTON 253 



towards and away from the sun. We do not see anything odd in 

 this distribution of comctary orbits; equally we may anticipate a 

 preferential motion of the same kind in the orbits of the stars. 



One doubt arose in connection with this suggestion which needed 

 theoretical examination. If the main lines of stellar traffic all 

 converge to the center of the system, will there not be a frightful 

 congestion at the center? It is true that the radial roads are laid 

 down as preferential only, but, allowing for the divergence of in- 

 dividual stars from the general direction, there would still seem to be 

 fear of a congestion in the inner part of the galaxy so great as to 

 lead to trouble. Mathematical calculation has, however, dispelled 

 this fear. One way of relieving congestion is to speed up the traffic 

 at the spot concerned. That applies to the stellar system ; the stars 

 which approach the center put on speed under the gravitational 

 attraction, and by clearing out of the neighborhood as quickly as 

 possible keep down the density to a tolerable value. Apparently, 

 then, nothing stands in the way of adopting Turner's suggestion 



But Kapteyn and others w^ho followed him pinned their faith to a 

 different theory. They believed that the line of star streaming was 

 transverse to the radius; this would mean that the stars tend espe- 

 cially to move in circular orbits about the center like planets, only 

 (unlike the planets) they go either way round indifferently. This 

 seems a very artificial disposition, and it could scarcely be regarded 

 as an explanation of star streaming in the sense that Professor Turn- 

 er's suggestion was. However, the first question is not which scheme 

 affords the better explanation but which gives a correct representa- 

 tion of the facts. Now that there is general agreement as to the di- 

 rection of the center, we can answer at once; the direction of star 

 streaming is radial, or nearly radial,*^ as Turner supposed. But the 

 observational decision in favor of radial star streaming is recent, and 

 theory had first innings in the contest Radial versus Transverse star 

 streaming. 



Adopting the specification of star streaming, introduced by 

 Schwarzschild, I was able to show rigorously that in a steady system 

 with star streaming the preferential motion must necessarily be 

 radial. Shortly afterwards J. H. Jeans showed rigorously that in a 

 steady system with star streaming the preferential motion must 

 necessarily be transverse.^ I doubt whether many spectators of the 

 game understood how to interpret the score of " one all " which was 

 the apparent result of the contest. Possibly they thought it was just 

 a case of two theorists contradicting one another as usual. But 



•The most trustworthy determinations give a difference of about 10* between the line 

 of star streaming and the radius ; it is difficult to decide whether the discrepancy is large 

 enough to be significant. 



' Unless the system la spherical— an exception Irrelevant to the study of our own 

 highly oblate galaxy. 



