247-250] Stellar Motions in the Galactic Universe 241 



stars (B to F4>) and second type stars (F5 to M) it is found that there are 

 723 second type stars of which 371 are receding while 352 are advancing, the 

 average velocity being found to be a velocity of recession of less than 1 km. 

 a second. Of the first type stars, however, 215 are receding while only 122 

 are advancing, and there is an average velocity of recession of 3'33 kms. a 

 second. Farther analysis shews that the motion of recession may be attributed 

 almost entirely to the 5-type stars, 138 of these shewing an average velocity 

 of recession of 4*93 kms. a second. 



Thus although stars of types A, F, G, K, M behave very approximately as 

 required by the steady-state law (576), it is quite clear that the -type stars 

 do not, at least if the most direct and obvious interpretation is put upon the 

 observations. Campbell and others are inclined to explain the observations 

 away by supposing that there may be a systematic error in the spectroscopic 

 determination of radial velocities, but our theory suggests another explanation 

 namely, that the universe is still expanding and that the _M-type stars, 

 being nearer to their steady state than the 5-type stars, are expanding less 

 rapidly. 



250. We have discussed five properties which ought to be observed in our 

 system if the final steady-state law (576) gave a tolerable approximation to 

 the motion. Of these properties we found at least a strong tendency for two 

 to be obeyed by all classes of stars. The remaining three were obeyed tolerably 

 well by all classes of stars except the $-type stars and possibly some of the 

 .A -type stars. 



Let us for the moment consider our universe as it would be if these latter 

 stars were blotted out of existence altogether. Then we have a universe of 

 which we can understand the mechanism well enough ; the motion is in 

 accordance with the laws of statistical mechanics and the system is exactly of 

 the type we should expect to find formed as the final product of a rotating 

 nebula. 



To a first approximation, to which the J/-stars conform particularly well, this 

 universe is simply a mass of stars rotating in an equilibrium configuration, the 

 angular velocity being everywhere uniform and of the order of 1" in 300 years. 

 Superposed on to the rotational velocity (which does not come into our 

 observational data, since the comparison stars share it with other stars) are 

 individual velocities of separate stars ; these, to our approximation, are dis- 

 tributed according to Maxwell's law, the mean velocity depending on a star's 

 mass but not on its position in space. The shape of this universe is what we 

 have called a pseudo-spheroid. The whole equilibrium is analogous to that 

 of a rotating mass of gas, the stars forming the " molecules." Throughout the 

 central portion, this " gas " is in isothermal equilibrium, although doubtless 

 this condition changes at the boundary. In accordance with the laws of 

 statistical mechanics, the heavier stars tend to congregate near the centre, 

 j.c. 16 



