APPENDIX. 



THE CALIBRATION OF THE CALORIMETER. 



The value 15.44 milligrams for the weight of mercury drawn into 

 the calorimeter by the introduction of one calorie of heat was checked 

 by measuring the heat of reaction of hydrochloric acid and sodium 

 hydroxide. The advantage of this method of calibration is that it 

 avoids all measurements of temperature and all error due to loss of 

 heat in dropping a hot body down the tube of the calorimeter. 



Half-normal solutions were used. The acid was carefully prepared 

 and standardized, both against solutions of known strength and by 

 determining the chlorine as silver chloride. Ten cubic centimeters 

 were used. The alkali was prepared with less care and used in 

 excess, the use of the ice calorimeter requiring that but one reagent 

 be present in known amount and high purity. The weight of mer- 

 cury drawn in in three runs was 1.1731, 1.1729, and 1.1 725 grams; 

 the average being 1.1728 grams. For the solution used (0.5078 nor- 

 mal) this gave 14,958 for the heat of neutralization. 



Such calibration is not entirely satisfactory because this value has 

 been determined before only at higher temperatures and the specific 

 heats of dilute solutions are not known with sufficient accuracy to 

 permit the calculation from one temperature to another. Julius 

 Thomsen J has studied the change of heat of neutralization with 

 temperature and gives values for this reaction at 25 C. and 10 C. 

 The value obtained appears high when compared with these two, 

 but it is to be noted that the accepted value for this heat of reaction 

 at 18 is low by the same comparison. In fact, the three values for 

 18, 10, and C. lie nearly in a straight line. It is hoped that the 

 heat of this reaction at zero will be accurately determined by other 

 methods and so remove the only objection to the use of this excellent 

 method of calibration. 



PROCEDURE IN GENERALIZING RESULTS. 



Upon completion of the measurements and calculation of the 

 results it becomes necessary to determine the best representative 

 value and to measure the consistency of the results, and, on the basis 

 of this measure, to form an estimate of the reliability of the repre- 

 sentative value determined. Accurate results are always discordant, 



i Thennochemlsche Untersuchungen. Bd. I, p. 64. 



19 



