ENERGY ACCELERATIONS. A STUDY IN ENERGY PARTITION AND 

 IRREVERSIRILITY 



BY 



G. H. BRYAN, Se D , F. R. S. 



lu tlie Pfiilosop/iical Magazine for JauuarV; Lord Rayleigh advan- 

 ces tlie view tliat tlie tinie-averages of the energy componeuts of a 

 siugle dynainical sjstem are distributed according to Maxwell's Law 

 of Partition provided tliat tlie sjstem satisiies certain conditions. Thèse 

 conditions appear to recpiire that tlie system s.hall at some time or 

 otlier pass tlirougli every possible distribution of coordiiiates and velo- 

 cities consistent with tlie équation of energy. 



It is certain that the inajority of Systems occurriug in ordinary dyna- 

 mics do not satisfy any such condition. Tt seems to me that our know- 

 ledge of the jn-operties of matter tends to make us disbelieve that such 

 a condition may be assumed as necessarily holding good in any Systems 

 of molécules except those Systems of ])erfectly elastic colliding sphères 

 or rigid bodies which are proved by Messrs Watson & Burbury and 

 others to represent the properties of an idéal perfect gas. The tendency 

 of atoms to forra chemical compouuds and to arrange theraselves in 

 crystalline groups, the phenomena of change of state, the coexistence 

 of a number différent phases ail appear to indicate that the molécules 

 of matter instead of passtng through ail possible contigurations consis- 

 tent with the energy-e([uation tend to distribute themselves about cer- 

 tain confîguratious and to entirely avoid other configurations. 



In the problem now considered, the probability that the coordinates 

 of a System shall lie between assigned limits is supposed to be given. 



