38 Reynolds and Moorby, Equivalent of Heat. 



delivered through a two-way tipping switch, either 

 waste or into a tank which stood on the platform of a weigh- 

 ing machine. When the engine was running steadily, the 

 discharge temperature could be regulated to the required 

 point by means of a valve inserted in the supply pipe 

 between the ice-cooler and first thermometer. If this 

 valve was opened, more water was admitted to the brake, 

 and consequently the temperature of discharge fell and vice 

 versa. By this means the temperature of discharge in 

 most of the trials rarely varied by as much as 2° from the 

 desired point, viz., 212°, and the mean temperature of 

 discharge, as obtained from 30 observations in each trial, 

 never varied by as much as 1° from 212°. 



The temperature of supply varied between 327° and 

 34'3° according as the temperature of the water in the 

 town's mains was lower or higher. In any one trial the 

 supply temperature was exceedingly steady, often not 

 varying through an interval of 0'i° through an hour's run. 



3. The various losses of heat. 



As it was likely that in a pair of trials the losses of 

 heat would not be exactly the same, and that, therefore, 

 some error would appear in the differences of work and 

 heat used in the determination of the equivalent, an 

 estimate was made of the losses of heat occurring by 

 different means in any trial. 



These losses were as follows : — 



(i) Loss by radiation : 



The total loss in a trial was assumed to be propor- 

 tional to the mean difference of temperature between the 

 water discharged and the air of the engine room, this 

 latter temperature being always taken from a thermometer 

 standing in a definite position. A quantity R, represent- 

 ing the loss by radiation per trial per unit difference of 

 temperature between the brake and the air was deter- 



