30 IN STARRY REALMS. 



source is not at all sufficient to counteract the direct loss 

 from radiation, and consequently the temperature falls and 

 the cooling of the body progresses. 



But now let us suppose the case of the cooling of a 

 globe of gas. No doubt the condition supposed is one 

 that we can hardly reproduce in our laboratories, but we 

 can imagine a globe of gas in space not enclosed in any 

 envelope, but merely taking the shape that is given by 

 the mutual attraction of its particles. We can indeed 

 see bodies apparently of this description in our telescopes, 

 and we call them planetary nebulae. Imagine a globe of 

 this kind to be dispensing its heat by radiation. In con- 

 formity with the general law, the loss of heat must be 

 accompanied by contraction. It is in the nature of a gas 

 to contract for a given fall of heat, through a much greater 

 range than that through which a solid contracts. The 

 consequence is that though the globe has undoubtedly 

 parted with some heat by radiation, yet the diminished 

 bulk of the contracted gas displays to such advantage the 

 heat still remaining, that the temperature of the mass 

 actually rises instead of falls ! 



This seems a proposition so remarkable that one may well 

 hesitate before accepting it. There can, however, be no 

 doubt of its truth, and it leads to a very remarkable result. 

 Suppose a great volume of gas were permitted to part 

 with what heat it had by radiation. If the volume were 

 large enough, and if the gaseous bodies of which it was 

 composed were of a nature that would admit of their 

 condensation, then the following is the course which 

 events would take. I am first supposing that the tempera- 

 ture of the gas is but little above that of the surround- 



