202 PROCEEDINGS OF THE AMERICAN ACADEMY 



hvdrate may not produce a basic precipitate as well as potassic hydrate, 

 and thus render both methods open to the same objection. 



However this may be, the results which I obtained soon led to the 

 suspicion that the reaction with potassic hydrate was not so simple as 

 is indicated by the above equation. One question after another 

 arising which I was unable to answer, I have been led step by step 

 to what may be called a thermochem'cal analysis of the reaction in 

 question. The first pait of the results obtained constitutes the sub- 

 stance of the following paper. At a later date, I hope to jmblish a 

 determination of the precipitation heat of aluminic hydrate by a 

 method free from these objections. The first series of experiments 

 involves the determination of the reaction heat of alum and potassic 

 hydrate, the proportion of the latter present varying from .5 to 10 

 molecules. This is followed by a parallel series determining the com- 

 position of the precipitates obtained in the calorimetric experiments. 

 Before considerinji in detail the results of the first series, I give a brief 

 description of the calorimetric methods and of the calculations that 

 were used. 



Method and Calculations. 



The general method of heat measurement is that described by 

 Berthelot. The calorimeter used was made after the specifications 

 given by him* without essential variation. The inner platinum calo- 

 rimeter has a capacity of somewhat more than 1,200 c.c. The stirrer 

 is a simple one, a piece of platinum foil fused to a glass rod. The ther- 

 mometers were made by Baudin of Paris especially for thermochemical 

 work after specifications by Berthelot, f and were calibrated by Pernet 

 at the Bureau International des Poids et Mesures. The zero points 

 have been recently redetermined. The three pairs of duplicates taken 

 together give a range of temperature from 0° to 33° C. They are 

 graduated to fiftieths of a degree, and the readings were made, usually 

 with the aid of a small hand lens, to ^vs of a degree, and often to ^J^- 

 The experiments were performed in a large basement room with- 

 out artificial heat, where, during a portion of the year at least, the 

 temperature conditions were fairly satisfactory. The operation con- 

 sisted, with but one or two exceptions, in mixing in the calorimeter 

 two portions of liquid of 500 c.c. For this measurement two flasks 

 were used, their graduation being verified by weight. 



Solutions containing the requisite quantity of constituents were pre- 

 pared, diluted to somewhat less than 500 c.c, and allowed to stand in 



* Essai de Me'canique Chimique, i. 140. t Ibid., p. 164. 



