ROTATION OF THE GALAXY EDDINGTON 255 



The coincidence was not unexpected so long as both were regarded as 

 " at rest," but it takes on a new significance when it is understood to 

 mean that both stars and cloud are moving about the center of the 

 galaxy with the same orbital velocity (of order 250 km per sec). 

 That they should agree to within 1 per cent is indeed surprising ; and 

 I think that the explanation of this coincidence should take preced- 

 ence over many of the other aims of dynamical theory. 



It seems impossible to admit any kind of interaction which would 

 tend to jdrag along the stars with the cloud or the cloud with the 

 stars. We have to regard them as two intermingled systems inde- 

 pendent in all respects except that the controlling gravitational field 

 is the same for both. If there is any community of motion it must 

 be because the same causes have operated in both and not because 

 one has constrained the other. At first sight a similarity seems 

 plausible. We often treat the stellar system as a glorified gas with 

 stars for molecules. Is it not then a case of two " gases " finding 

 their own conditions for equilibrium independently? Why should 

 we be surprised that they both hit on the same solution ? Neverthe- 

 less, I admit that I am surprised; the reason is that the two gases 

 differ enormously in viscosity. 



An atom in the cosmic cloud may in the course of its wanderings 

 expect a collision with another atom about once a year ; in that time 

 it traverses a path about equal to the distance of the earth from the 

 sun. This is a long free path according to ordinary standards, but 

 it is insignificant in the scale of the stellar universe. On the other 

 hand the free path of a star is practically infinite ; it can go hundreds 

 of times around its orbit from one side of the galaxy to the other 

 without appreciable risk of deflection. The length of the free path 

 determines the viscosity of a gas. The viscosity of the cosmic cloud 

 is negligible for astronomical purposes; the viscosity of the star-gas 

 is enormous. In fact the stellar universe, regarded as a gas, is the 

 stickiest thing you could possibly imagine. 



The theory of a rotating nonviscous gas is familiar, and it can be 

 applied directly to the cosmic cloud. We can at once derive an 

 important result; the motion of the cloud must correspond almost 

 exactly to that of a particle revolving in a circular orbit about the 

 center. A slight deviation — if only a few km per sec. — would set 

 up an enormous density gradient in the cloud, so that in one or other 

 direction the stars would be embedded in a solid jam of cloud-atoms.* 



" I ought to mentlou that the tinio of relaxation toward dynamical equilibrium is ]onger 

 for the cloud than for the stars owing to the lower velocity of the particles. We may say 

 that the velocity of sound in the cloud (about 3 km per sec.) is less than the velocity of 

 sound in the star-gas ; and accordingly the pressure-waves which level out the distribution 

 of cloud matter take a longer time to travel through the galaxy than those which level 

 out the distribution of stars. It is, therefore, possible that if we adopt a short-time scale 

 the cosmic cloud may not yet have reached the dynamical equilibrium assumed in my 

 discussion. 



