92 Proceedings of the Royal Societ)j of Victoria. 



if it leave the body its entropy falls ; to this rule there ai^e, and 

 can be, no exceptions. 



There is no lack, then, of definiteness about this property ; 

 the difficulty in grasping its signiticauce is due to its lack of 

 familiarity. AVe are accustomed to think and speak of volumes 

 and temperatures ; probably not one educated man in five could 

 define either accurately, but all are familiar with the terms. 

 The reason for this is, in my opinion, to be found in the fact 

 that instruments for measuring both quantities are in common 

 every-day use; but, so far, we have had no instrument for 

 measuring entropy. The quantity is therefore an unfamiliar 

 one ; and the fact that it can easily be calculated from measure- 

 ments of other quantities helps the student very little in his 

 efforts to gain a clear idea of it. 



These considerations — combined with the fact that the instru- 

 ment, once constructed, would probably be useful — have induced 

 me, for a good many years past, to search for a mode of 

 measuring entropy changes on the scale of a suitable instrument. 

 So far I have not succeeded in constructing a single apparatus 

 which would do this in all possible cases, though I have not 

 yet given up all hope of ultimate success in this direction. 

 The difficulty of the investigation lies in the fact that the 

 change which takes place in the entropy of a body during a 

 transformation is expressed mathematically as the sum of two 

 more or less independent integrals ; I do not yet see how an 

 instrument is to be made which will evaluate these simulta- 

 neously and add them together. 



This difficulty is avoided if we restrict the generality of the 

 apparatus. The restriction is by no means such a disadvantage 

 as it may appear at first sight, seeing that, in the great majority 

 of the transformations we meet with in practice, one or other of 

 the three quantities on which the measurement of the entropy 

 depends is kept constant ; only one instrument is then required 

 for observation of the entropy change. 



If we design instruments to measure entropy changes at 

 constant pressure, constant volume, and constant temperature, 

 three will be required ; one will do the measurements both at 

 constant pressure and constant volume for either solids, liquids, 

 or gases ; but two will be required for constant temperature, one 



