Comments On The Note By E. H. Kennard On Entropy, Reversible Processes And Thermo-couples

242 PHYSICS: P. W. BRIDGMAN PROC. N. A. S. COMMENTS ON THE NOTE B Y E. H. KENNARD ON "ENTROPY, RE VERSIBLE PROCESSES AND THERMO-CO UPLES" By P. W. BRIDGMAN JEFFERSON PHYSICAL LABORATORY, HARVARD UNIVERSITY Communicated February 9, 1932 The matter discussed by Professor Kennard in the preceding note is of sufficient importance to justify an attempt to make somewhat plainer my own point of view and to indicate what I believe still remain the most important differences between our attitudes after a personal discussion which has now extended over several years and which I am sure both of us regard as profitable. In the first place, I must correct the impression, justified it is true by the form of presentation of my previous paper,' that I claimed to have a rigorous proof of Kelvin's thermoelectric relations from the assumptions of classical thermodynamics alone. An examination of the argument there presented will show that something had to be assumed in addition to the classical thermodynamics of Kelvin and Clausius. What I assumed was the existence in nature of intrinsically irreversible processes which are accompanied by definite and characteristic entropy increases whenever they occur, irrespective of other processes which may or may not be occurring simultaneously. Examples of such intrinsically irreversible processes are heat conduction and generation of Joulean heat when an electric current flows against resistance. In particular, whe'n an amount of heat Q flows by conduction through a temperature difference AT at a mean temperature T the entropy increase is QAT/T2. In most situations presented by classical thermodynamics the consideration of such intrinsically irreversible processes is not necessary because by a suitable arrangement of the proportions of the system the relative effect of the irreversible processes may be made vanishingly small. In t