112 THE AIM AND ACHIEVEMENTS OF SCIENTIFIC METHOD. 



and C, and may be taken therefore as measures of the 

 differences between these temperatures and that of A. Thus, 

 if the second engine does twice as much work as the 

 former, the difference of temperature between A and C may 

 conveniently be taken to be twice as great as in the case of 

 A and B. We may assign, then, any arbitrary number to the 

 temperature of A, and in this way correlate with any standard 

 temperatures available (such as the boiling and melting points 

 of pure substances), numbers which will be independent of the 

 properties of any particular substance. It will be convenient, 

 however, so to choose the number assigned to the temperature 

 of A that in the ideal case in which all the heat drawn from 

 the source is converted into work, the number assigned to the 

 temperature of the condenser shall be zero.* 



53. 



Unfortunately, however, this correlation of temperatures 

 with quantities of mechanical work does not enable us to 

 dispense with the notion of temperature as a definite state of 

 a body which is the object of our sensations of hotness or 

 coldness. The argument which made this unambiguous corre- 

 lation possible is not developed from mechanical data only, 

 even if we ignore the difference between heat and kinetic 

 energy. It involves our direct experience (formulated in the 

 axioms of Clausius and Lord Kelvin) that hot bodies in the 

 presence of colder ones get cooler, never hotter, except 

 (ultimately) by the intervention of animate agencies. f 



To effect the complete annexation and assimilation of the 

 science of Heat, Dynamics must, then, give some account of 

 the behaviour of solid and liquid bodies from which (as in 

 the Kinetic Theory of Gases) this irreversible property of 

 temperature can be deduced. 



* The "absolute "zero. 

 t Preston, op. cit., p. 612. 



