272 Methods of determining and calculating 



the temperature of the water, the latter receiving from 1000 to 1200 

 degrees of heat, for every unit of water condensed, will cause an er- 

 ror in excess. If however the vessel have remained in its cold state 

 for some time, and then received a considerable elevation of tempe- 

 rature from the hot body, the whole exterior of the vessel will act 

 as the wet bulb of a thermometer, and tend to keep the temperature 

 of itself and its contents down to the evaporating point. This would 

 cause a serious error in defect. Both these errors are obviously to 

 be avoided by not allowing the temperature of the vessel to sink be- 

 low the dew point. In regard to the relative temperatures of the 

 vessel and the surrounding air, we must observe that as the latter 

 part of the process, when the solid and the water are approaching an 

 equilibrium, goes on very slowly, it will be necessary to commence 

 our experiment with the water nearly as much below the actual tem- 

 perature of the apartment as the increase of temperature is expected 

 to bej in order to terminate as little as may be above the surrounding 

 air. These two conditions of beginning above the dew point and 

 never ending much above the temperature of the air can be compli- 

 ed with only when the air is tolerably dry. Such should therefore 

 be the state of weather selected for experiments of this nature. 



4. The construction, magnitude, and specific heat of the ther- 

 mometer, used to measure the temperature of the water, is an object 

 of some consequence in the determination of this delicate question. 

 To carry entire accuracy into the subject it will be necessary to 

 know the separate weights of the materials which compose it, and 

 their several specific heats, and further to allow for an amount of wa- 

 ter precisely equivalent to that part of the thermometer which is 

 immersed during the experiment. In obtaining a thermometer for 

 this purpose I caused the tube to be carefully measured and weighed 

 before the bulb was blown, to ascertain its weight per inch in length, 

 then knowing the length used to form the bulb it was easy to ascer- 

 tain the number of grains of glass immersed in any given experiment. 

 By again weighing after the thermometer was filled, the weight of 

 mercury it contained was exactly known, and by weighing the scale 

 separately and knowing its specific heat, the equivalent in water was 

 found answering to any portion of the whole instrument, which may 

 be entered along the scale near the thermometric degrees. The ne- 

 cessity of allowing for a scale may however be obviated by using a 

 naked-bulb thermometer provided the range be sufficient without in- 

 cluding the naked part of the stem. But to attain this end and at 



