56 CORRELATION OF PHYSICAL FORCES. 



up-and-down stroke of a piston, we must, as I have already 

 stated, heat and cool, just as with a magnetic machine we 

 must magnetise and demagnetise in order to produce a con- 

 tinuous mechanical effect ; and although, from the impos- 

 sibility of insulating heat, some heat is apparently lost in the 

 process, the result may be said to be effected by the transfer 

 of heat from the hot to the cold body, from the furnace to 

 the condenser. But we may equally well say that the heat 

 has been converted into mechanical force, and the mechanical 

 force back into heat ; the effects are always correlative, as 

 are the mechanical effects of an air-pump, with which, as 

 we dilate the air on one side, we condense it on the other ; 

 and as we cannot dilate without the reciprocal condensation, 

 so we cannot heat without the reciprocal cooling, or vice versa. 



Hitherto the resistances of the piston or of any superim- 

 posed weight have been thrown out of consideration, or, what 

 amounts to the same thing, it has been assumed that the 

 weight raised by the piston has descended with it. The heat 

 has not merely been employed in dilating the air or vapour, 

 but in raising the piston with its weight If, as the vapour 

 is cooled, the weight be permitted to descend, its mechanical 

 force restores the heat lost by the dilatation ; but in this case 

 no part of the power has been abstracted so as to be em- 

 ployed for any practical purpose ; this question then follows, 

 what takes place with regard to the initial heat, if, after the 

 ascent of the piston, the weight be removed so as not to help 

 the piston in its descent, but to fall upon a lever or produce 

 some extraneous mechanical effect ? 



To answer this question, let us suppose a weight to rest on 

 a piston which confines air at a definite temperature, say for 

 example 50, in a cylinder, the whole being assumed to be 

 absolutely non-conducting for heat. A part of the heat of 

 this confined air will be due to the pressure, since, as we have 

 seen, compression of an elastic fluid produces heat. 



Suppose, now, the confined air to be heated to 70, the 

 piston with its superincumbent weight will ascend, and the 

 temperature, in consequence of the dilatation of the air, will 

 be somewhat lowered, say to 69 (we will assume, for the sake 



