1KS2.] odJ [Eddy. 



present) proved, but it, so far, has never been found in disagreement with 

 experience. 



It is well known that Maxwell has proposed a process to accomplish this 

 very object, namely to transfer heat from a colder to a hotter body, in the 

 following manner : If we suppose minute beings, endowed with senses 

 suliiciently acute and having a corresponding agilitj% to guard minute 

 openings in the diaphragm separating two portions of the same gas, 

 which openings are only large enough for a single molecule to pass at 

 once, they would be able without expenditure of energy to open and close 

 the openings in such a way as to allow each molecule impinging at an 

 opening to pass through or not, as they should choose. If they permitted 

 only those molecules having more than the mean vis viva to pass in one 

 direction and only those having less than the mean to pass in the opposite 

 direction, then the gas on one 'side of the diaphragm would gain energy 

 at the expense of that on the other side. That this process is actually at 

 present beyond human ability does not show that we may not at some 

 future time be able to accomplish what Maxwell proposed. If this be 

 admitted, then the conclusions which I shall draw later from lack of gen- 

 erality in the second law of thermodynamics flow to a limited extent from 

 the possibility of this process. 



But Maxwell's process assumes the kinetic theory of gases as its basis, 

 and stands or falls with it. 



And if the second law is a necessary ultimate mechanical principle, 

 holding for all bodies great and small, the above consequence of the 

 kinetic theory of gases being in contradiction to the secorld law is fatal to 

 the validity of the kinetic theory. But I do not now so regard the second 

 law. I am compelled to regard it as merely an approximation in the case 

 of radiations, and to regard it in general, with Maxwell and with Boltz- 

 mann,* as merely the mean result flowing from the laws of probability ; 

 though it had previously seemed to me possible to show it to depend upon 

 fundamental considerations respecting the nature of heat as a forna of 

 energy, as was stated in my work previously referred to. 



To avert to the consequences which are thus made to flow from the es- 

 tablished fact of the finite velocity of radiant heat, we may mention that 

 if the law of the dissipation of energy is no longer to be regarded as of 

 universal validity, it being obviated by the process of the syren, it is just 

 as possible to avail ourselves of the heat stored in cold bodies as in hot 

 ones, and thus to employ the heat of a glacier to drive a steam engine, or 

 to perform other like feats heretofore regarded as impossibilities. "When 

 I say it is just as possible, I do not imply that it is now just as practicable, 

 or perhaps ever will be so. 



That these observations are just, is seen when w^e reflect that the pro- 

 cess of the syren simply heats a given body at the expense of any othei', 

 regardless of temperatures, by a method requiring the expenditure of no 



*Weir. Sltzb. Band. Ixxvi, Ixxvili. 



PROC. AMEK. PHILOS. SOC. XX. 112. 2q. PRINTED AUGUST 14, 1882. 



