166 PROFESSOR W. THOMSON ON THE SECULAR COOLING OF THE EARTH. 



21. In the process of refrigeration, the fluid must (as I have remarked regard- 

 ing the sun, in a recent article in " Macmillan's Magazine,"* and regarding the 

 earth's atmosphere, in a communication to the Literary and Philosophical Society 

 of Manchester!) be brought by convection, to fulfil a definite law of distribution 

 of temperature which I have called " convective equilibrium of temperature." 

 That is to say, the temperatures at different parts in the interior must differ 

 according to the different pressures by the difference of temperatures which any 

 one portion of the liquid would present, if given at the temperature and pressure 

 of any part, and then subjected to variation of pressure, but prevented from 

 losing or gaining heat. The reason for this is the extreme slowness of true 

 thermal conduction ; and the consequently preponderating influence of great 

 currents throughout a continuous fluid mass, in determining the distribution of 

 temperature through the whole. 



22. The thermo-dynamic law connecting temperature and pressure in a fluid 

 mass, not allowed to lose or gain heat, investigated theoretically, and experimen- 

 tally verified in the cases of air and water, by Dr Joule and myself, t shows, 

 therefore, that the temperature in the liquid will increase from the surface down- 

 wards, if, as is most probably the case, the liquid contracts in cooling. On the 

 other hand, if the liquid, like water near its freezing-point, expanded in cooling, 

 the temperature, according to the convective and thermo-dynamic laws just 

 stated (§§ 21, 22), would actually be lower at great depths than near the sur- 

 face, even although the liquid is cooling from the surface ; but there would be a 

 very thin superficial layer of lighter and cooler liquid, losing heat by true conduc- 

 tion, until solidification at the surface would commence. 



23. Again, according to the thermo-dynamic law of freezing, investigated by 

 my brother,§ Professor James Thomson, and verified by myself experimentally for 

 water, || the temperature of solidification will, at great depths, because of the gi'eat 

 pressure, be higher there than at the surface if the fluid contracts, or lower than 

 at the surface if it expands, in becoming solid. 



24. How the temperature of solidification, for any pressure, may be related 

 to the corresponding temperature of fluid convective equilibrium, it is impossible 

 to say, without knowledge, which we do not yet possess, regarding the expansion 



* March 1862. 



f " Proceedings," Jan, 1862. " On tlie Convective Equilibrium of Temperature in the 

 Atmosphere." 



J Joule, " On the Changes of Temperature produced by the Rarefaction and Condensation of 

 Air," Phil. Mag. about 1844, Thomson, " On a Method for Determining Experimentally the Heat 

 evolved by the Compression of Air ; Dynamical Theory of Heat, Part IV.," Trans. R. S. E., Session 

 1850-51 ; and reprinted, Phil. Mag. Joule and Thomson, " On the Thermal Effects of Fluids in 

 Motion," Trans. R. S. Lond., June 1853 and June 1854. Joule and Thomson, " On the Alterations 

 of Temperature accompanying Changes of Pressure in Fluids," Proceedings R. S. Lond., June 1857. 



§ " Theoretical Considerations Regarding the Effect of Pressure in lowering the Freezing- Point 

 of Water," Trans. R. S. E., Jan. 1849. 



II Proceedings R. S. E., Session 1849-50. 



