120 KEPORT — 1891. 



respecting the permanent distribution of energy in a system of material 

 points under intermolecular forces, it would be premature to form con- 

 clusions regarding the tendency towards the equalisation of energy, except 

 in those cases where the only reactions between the points are those due 

 to impact. 



50. If we regard the whole matter as one of probabilities, the argu- 

 ment derived from reversing the system may be met without an appeal 

 to tiie luminiferous sether. Although a conservative dynamical system 

 is always reversible, the reversed motion may not unfrequently be 

 dynamically unstable in the highest degree. One of the best illustrations 

 in point is afforded by the impossibility of riding a bicycle backwards 

 (i.e. with the steering wheel behind) ; here the forward motion is stable, 

 but the reversed motion is highly unstable. 



Take, then, a system of material points or colliding spheres all tend- 

 ing towards the ' special state.' If the motion is slightly disturbed they 

 will still tend towards the ' special state,' and the effect of the disturb- 

 ance in modifying the character of the motion will diminish without 

 limit. But if we suppose at any stage of the process that the motion 

 of every point is exactly reversed, then the difference between the dis- 

 turbed and undisturbed reversed motions will increase without limit, and 

 the disturbed reversed motion will tend towards a very different state from 

 that from which we started. In a very short time we shall have entirely 

 different series of collisions taking place in the disturbed and undisturbed 

 reversed motions. When, therefore, we consider the immense number of 

 molecules present in any body of finite size, it is not hard to understand 

 that the probability of the energy tending towards an unequal distribu- 

 tion is infinitesimally small, for just the same reason that if any two 

 different substances in a minute state of subdivision have become 

 thoroughly mixed it is impossible to separate them again by simply 

 stirring them up. There is nothing inconceivable about such a separa- 

 tion, but the chances are so overwhelmingly against it that we may with 

 absolute certainty declare the separation impossible. In this manner 

 there is no difhculty in understanding how on statistical grounds alone 

 we may be able to state with absolute certainty that ' heat cannot pass of 

 itself from a cold body to a hot body.' 



Of course evidence of this kind is speculative, and, moreover, only 

 affords a possible explanation, and not a proof, of the principle of 

 degradation of energy. 



But, as it has been necessary to suppose space furnished throughout 

 with an aether in order to account for electrical and optical phenomena, 

 allowance must be made for the fact that this iether will in all probability 

 play a pi-omiuent part in thermal phenomena, more especially as it is the 

 medium by which radiant heat is propagated. The great velocity of 

 light shows that the ajther can have but a very small capacity for radiant 

 energy, and, therefore, that its presence will not materially affect the 

 results of investigations relating to reversible thermodynamic processes, 

 while it will certainly facilitate the dissipation of energy. It must not, 

 however, be thought that researches relating to heat are worthless be- 

 cause they do not take account of the aether ; for do not such researches 

 fulfil what should be the highest object of scientific enquiry — namely, of 

 helping us to 'judge the unknown from the known ' ? 



