82 FIRST SERIES ON CONDENSATION. 



of temperature less than the tenth part of a degree (one 

 twentieth of an inch on Joule's thermometer), which were 

 correct to the two-hundredth part of a degree, as is shown 

 by the results deduced from them. 



With these appliances Joule makes three series of 

 experiments. In the first he pumps his carefully-dried air 

 slowly into a receiver, then determines the increase of 

 temperature in the surrounding water. This increase is, he 

 considers, due to the friction of his pump, to radiation 

 from the room, as well as to the work done in compressing 

 the air. He therefore makes experiments, from which he 

 determines the amount of the rise due to the first two 

 causes, and subtracting this he gets the rise of temperature 

 0°*345 Fah. due to the last cause. This rise, multiplied 

 by the capacity for heat of the water, gives him the heat 

 resulting from the condensation of the air. He has now to 

 obtain the work which has been done in compressing the 

 gas. This he does, not by measuring the resistance and 

 motion of the pump, but by measuring the volume of the 

 compressed air when expanded again to the atmospheric 

 pressure. This is 21*654 times the volume of his receiver, 

 which he has calibrated. Then, by Boyle's law, he estimates 

 that the final pressure of the air in his receiver is 21*654 

 times the pressure of the atmosphere at the time, and by 

 this means and the hyperbolic relation between the pressure 

 and volume he estimates the work done. This process 

 would be strictly accurate, provided the air in the pump 

 had maintained a constant temperature, which it would not 

 have done had not sufficient time been allowed for the heat 

 resulting from condensation to pass from the air into the 

 water during the intervals of compression. Joule nowhere 

 in his paper mentions the necessity of this condition ; but 



