OF ARTS AND SCIENCES. 209 



The method pursued was the following : 50 c.c. of alum solutiou, 

 containing five grammes of the crystallized salt, were made up to a 

 little less than 125 c.c. The required quantity of KOII solution was 

 also made up to about the same volume. These two solutions were 

 then thoroughly mixed, the volume of the mixture made up accurately 

 to 250 c.c. (2' 20° C), again well shaken, and thrown upon a dry 

 filter, the filtration being hastened in some instances by using the 

 pump. From the filtrate separate portions of 50 c.c. each were taken, 

 in which were made duplicate determinations of SOg and AljOg, and 

 in several instances of K^O. The SO3 was determined, of course, by 

 precipitation with BaCl, ; the AI2O3, by precipitation with NH3 ; and 

 the K^O was weighed as sulphate after the removal of AI2O3. The 

 proportion of constituents and water is the same as in the calorimetric 

 experiments, the scale being simply reduced to one fourth. 



Analyses somewhat similar have been made by S.U.Pickering* 

 of the precipitates obtained from aluminic sulphate by addition of 

 varying quantities of NaaCOg. His method was to determine the SO3 

 and AloOg in the precipitate itself after thorough washing, although he 

 recognized the fact that its composition was changed by the operation. 

 This would be a serious objection to the use of this method for the 

 purposes of this investigation, although less so in his experiments under- 

 taken from a different standpoint. Our results, so far as they cover 

 the same points, do not conflict in the general direction of their indi- 

 cations, although quantitatively they differ somewhat. 



Mills also has determined simply the amount of ALOg precipitated 

 from alum solution by varying quantities of Na^COg-t 



It should be noted that the method of this investigation gives no 

 evidence as to the degree of hydration of the precipitate ; that point 

 I am obliged to pass by. 



In the first experiment of this series (Exp. A) the method was 

 applied in blank to a solution of alum, no KOH being added. The 

 results obtained were : 



AIO3 = .1080 grm. SO. = .3321 grm. 



= .1074 ^ = .3326 



Mean = .1077 Mean = .3324 



As these correspond to 1 grm. alum taken, multiplying by 100 con- 

 verts them into per cents. Correcting for the volume of pipettes and 



* Cliemical News, xlv. 121, 133, 146. 

 t Journal of Cliem. Society, [2,] xli. 341. 



VOL. XXI. (n. 8. XIII.) 14 



