208 PROCEEDINGS OF THE AMERICAN ACADEMY 



then six times their quantities, the volume 1,000 c.c. remaining the 

 same throughout. In the last case, the proportion is about the same 

 as in Thomsen's experiment. The data are shown in Table IV. At 

 the degree of concentration which corresponds to Exp. 36-39, the 

 formation heat of^K2S^4^^ found to be (Exp. 57-61) 14.63 Cal., 

 which agrees with ^ in Exp. 36-39 ; while for one sixth this dilution 

 i/is found to be 31.26 or 15.63 for h KoSO^ ; Exps. 53, 54, correspond 

 to Thomsen's conditions. 



This proves that the second half of the 9th raol. and the 10th mol. 

 in Series I. are not in chemical combination, or, at least, that thei'e is 

 no evidence of it in the thermal phenomena. In other words, the heat 

 disturbance ceases when 8.13 mol. KOH have been added, as indicated 

 in Line I., Plate I. 



As dilution was found to affect the determination of the formation 

 heat of I^SO^, it became desirable to test its influence upon the re- 

 sultant ff of Series I., in which this formation heat is the principal 

 element. To do this the reaction heat of 6 mol. KOH on I of alum 

 was determined, the absolute amount of substance undergoing change 

 in the standard volume, 1,000 c.c, being only one fourth the quantity- 

 used in Exp. 23-26, Series I. These data are recorded in Table V. 

 The mean result of the five experiments is 29.26 Cal., in place of 30.66 

 by the previous experiments. 



It seems not improbable, in view of these facts, that the tendency 

 of the Line III., Plate I., to rise between x = 1.5 and x = 5.1, may 

 be due to the increase in the amount of substance undergoing change 

 in the unit of volume. 



The complexity of the thermochemical results thus far described 

 made evident the necessity of a sharper definition of the chemical 

 changes actually taking place in the calorimeter. The study of these 

 constitutes the second part of this paper. 



PART II. 



General Method. — It would seem that the most direct method of 

 obtaining evidence as to the exact chemical change would be to deter- 

 mine the composition of the various precipitates produced in the 

 calorimetric experiments. The impossibility, however, of washing 

 them without changing their composition, indicated rather the analysis 

 of the filtrate from them as the preferable method. Either one deter- 

 mines only the total change in the distribution of constituents between 

 the soluble and insoluble portions of the system. 



