372 Mr. Joule mi the Changes of Temperature produced by 



thermometer had to be reduced in every instance, a ch'Cum- 

 stance which accounts for my having given the temperatures 

 in the tables to three places of decimals. 



It vk'as important to employ, for the purpose of containing 

 the water, a vessel as impermeable to heat as possible. With 

 this view, two jars of tinned iron, one of them every way an 

 inch smaller than the other, having been provided, the smaller 

 jar was placed within the larger one, and the interstice between 

 the two was closed hermetically. By this means a stratum of 

 air of nearly the same temperature as the water was kept in 

 contact with the sides and bottom of the inner jar. The jars 

 used in the other experiments which I shall bring forward 

 were constructed in a similar manner. Among other precau- 

 tions to ensure accuracy, proper screens were placed between 

 the vessels of water and the experimenter. 



My first experiments were conducted in the following man- 

 ner: — The pump and copper receiver were immersed into 

 45 lbs. 3 ozs. of water, into which the very sensible thermometer 

 above described was then placed ; whilst two other thermome- 

 ters were employed in order to ascertain the temperature of 

 the room and that of the water contained by the vessel W. 

 Having stirred the water thoroughly, its temperature was 

 carefully read off. The pump was then worked at a mode- 

 rate degree of speed until about twenty-two atmospheres of 

 air, dried by being passed through the vessel G full of small 

 pieces of chloride of calcium, were compressed into the copper 

 receiver. After this operation (which occupied from fifteen 

 to twenty minutes) the water was stirred for five minutes so as 

 to diffuse the heat equably through every part, and then its 

 temperature was again read off". 



The increase of temperature thus observed was owing partly 

 to the condensation of the air, and partly also to the friction 

 of the pump and the motion of the water during the process 

 of stirring. To estimate the value of the latter sources of heat, 

 the air-pipe A was closed, and the pump was worked at the 

 same velocity and for the same time as before, and the water 

 was afterwards stirred precisely as in the first instance. The 

 consequent increase of temperature indicated heat due to fric- 

 tion, &c. 



The jar was now removed, and the receiver having been 

 immersed into a pneumatic trough, the quantity of air which 

 had been compressed into it was measured in the usual man- 

 ner, and then corrected for the force of vapour, &c. The 

 result, added to 136'5 cubic inches, the quantity contained by 

 the receiver at first, gave the whole quantity of compressed 

 air. 



