76 On the Gas-Engine Indicator-Diagram. 



We are now engaged in an investigation involving much 

 higher differences of temperature than those of MM. Dulong 

 and Petit; and in so far as we have obtained results for tem- 

 peratures extending from about 800° 0. to 1300° C, we have 

 confirmed the conclusions of these gentlemen. Thus the loss 

 of heat per second at 776° C. being 97*2, the loss at 1292° C. 

 is 253-2. 



Our method of experimenting is, we believe, new. The 

 heated body is a spiral of platinum, whose change of electric 

 resistance when there is change of temperature is known. It 

 is surrounded by a vessel, blackened inside and maintained at 

 constant temperature. An electric current is made to pass 

 through the platinum spiral, maintaining it at any required 

 temperature. An ammeter and voltmeter enable the current 

 A and the difference of potentials V between the ends of the 

 platinum spiral to be measured. Then YA is the rate at 

 which heat is being radiated from the platinum, and V/ A is 

 the electric resistance- of the platinum, from which its tempe- 

 rature is known. We intend to investigate the influence of 

 high pressures of air and other gases. 



We are somewhat doubtful as to the weight which we 

 ought to give to the results obtainable from these experiments 

 on the loss of heat by solid bodies, since in our gas-engine 

 investigations, we deal with a mixture of hot gases ; and in 

 adopting the law of simple proportionality to difference of 

 temperature, we have been influenced by the fact that rate of 

 loss of heat by the fluid in the compression-stroke increases 

 much more slowly than is indicated in Dulong and Petit's 

 law, although the pressure of the fluid is increasing as well as 

 its temperature. The result given in § 11 is to the effect that 

 during compression the rate of loss of heat by the fluid is 

 nearly proportional to the -J§ power of the absolute tempera- 

 ture of the fluid, or to 



(0 + 333) if , 



if 6 is the difference of temperature from 60° C, the tempe- 

 rature taken as that of the cold cylinder. If q is this rate of 



loss, it is obvious that ~tq diminishes as increases. It will 



be observed that the probability of the piston's having a high 

 temperature causes this result to be even more curious than it 

 might otherwise appear to be. 



