660 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Previous Direct Methods of Measuring the Specific Heats 

 OF Salt Solutions. 



The brief discussion of tlie methods which previous investigators have 

 used in measuring the specific heats of salt solutions will serve both to 

 emphasize the most obstinate difTiculties in these measurements and to 

 furnish the basis for a more intelligent criticism of the new methods 

 to be described. 



Person,* in 1851, made a few rather inaccurate determinations in this 

 field. Later, Schullerf carried out a somewhat more extended series of 

 measurements. Both these investigators used the simple method, which 

 Andrfews had devised in 1845, for the determination of the specific heats 

 of liquids. A weighed amount of water was inclosed in a glass vessel, 

 into which a thermometer fitted. This served as a " calorifer." It was 

 heated to a given temperature, and then plunged into the solution and 

 used as a stirrer. The temperature of the solution was measured at fre- 

 quent intervals until it had become constant, or was regularly falling, and 

 the procedure of Regnault emploj^ed to correct for the radiation of heat. 

 The method of Andrews, as they employed it, had little to recommend 

 it beyond simplicity. The correction for radiation was uncertain, wliile 

 it is quite impossible for the thermometer in the calorifer accurately to 

 measure its temperature. 



Julius Thomsen t was the first to make a protracted study of the 

 specific heats of solutions. He pointed out the inaccuracies in the method 

 just montioned, and elaborated a very original method to overcome them. 

 A known quantity of heat was furnished to the solutions by burning suc- 

 cessively equal measured quantities of hydrogen in a jilatinura vessel, 

 which they successively surrounded. Since the pressure of the hydrogen 

 was kept accurately constant by an automatic device, the heating effect 

 was directly proportional to the time, and therefore by choosing an 

 initial temperature as much below the room temperature as the final was 

 above it, no correction for radiation was necessary. The burning hydro- 

 gen is certainly superior to the calorifer of Person and SchuUer ; but the 

 necessary apparatus is cumbrous, and the accuracy of the metliod is 

 limited by the accuracy of the volumetric measurement of the hydrogen. 

 The doubtful accuracy of the correction for radiation sets a further limit 

 to the refinement of the method. In the skilled hands of Thomsen, 



* Person, Ann. de Cliim. et de Phys. (3), 33, 437 (1851). 



t Scliiiller, Pogg. Ann., 136, 70 (18G9). 



t Tliomsen, Thermocliemische Untersucliungen (1882). 



