512 Mr. J. Parker on an Extension 



an exaggerated degree in measuring the temperature of a 

 couple of which one element is steel, by means of a second 

 couple placed inside an enclosure side by side with the steel 

 couple. The accession of heat in the enclosure due to the 

 reglow of the steel in the thermoelectric couple is not enough 

 to produce an effect on the thermometric couple. 



There is another peculiarity in the thermoelectric properties 

 of steel, to which Mr. Tomlinson refers (Phil. Mag. Jan. 1888). 

 A piece of steel wire is connected up with a galvanometer, 

 and is heated by a flame at a point remote from the galvano- 

 meter connexions. There are no signs of a resultant E.M.F. 

 in the circuit, even when the flame is moved, until 'darkening' 

 has occurred. Then, however, if the flame be moved along 

 the wire so that reglow takes place behind it in the cooling 

 part, and darkening in front in the part being heated, an 

 E.M.F. results, and persists only so long as the flame is kept 

 moving. 



In proving the rise of temperature throughout the mass of 

 steel during recalescence, I have shown that the temperature 

 of recalescence is different from that of darkening, the latter 

 being higher. This difference of temperature of what may be 

 regarded as the junctions of the hot altered steel with the cold 

 steel on each side, determines the E.M.F. in the circuit. 



LXIII. On an Extension o/" Carnot's Theorem. By J. Parker, 

 B.A., late Scholar of St. Johns College, Cambridge* . 



IT is well known that no engine can be more efficient in 

 transforming heat into mechanical work than Carnot's 

 reversible engine, and that all reversible engines are of the 

 same efficiency. It is also generally believed, without, how- 

 ever, being proved, that no irreversible engine can be so 

 efficient as Carnot's. This is the point we propose to examine. 



One of the principal conditions imposed on a system sub- 

 jected to a reversible operation is that, at every instant, the 

 system is in mechanical and thermal equilibrium. This con- 

 dition, however, also belongs to a certain class of irreversible 

 processes, as we shall presently see. 



Thus, when a saturated solution of sodium chloride is heated 

 with an excess of the salt, the salt is dissolved ; but if the 

 temperature is again reduced, the solution deposits, not the 

 anhydrous salt, but crystals of the bihydrate. Again, a salt 

 is neutral to pure ice at the freezing temperature of its crvo- 

 hydrate; and if heat is slowly imparted, the liquid cry ohyd rate 



* Communicated by the Author. 



