• OF ARTS AND SCIENCES. 215 



one of alum ; the precipitate then becomes permanent ; it is light yellow 

 in color, and has a gelatinous appearance which suggests an excess 

 of alumina. The reaction is evidently the same as that above given 

 for the precipitation of chromate of chromic oxide. 5 (KO CrOs) 

 + AlA 3 SO3, KO S03 = 4 (KO SO3) + 2 (KO 2 CrOs) + AI2O3 

 CrOs. The precipitate was drained, pressed between folds of filter- 

 paper under heavy weights, and air-dried. The filtrate, when evap- 

 orated and crystallized, gave crystals of bichromate of potash, and of 

 sulphate of potash mixed with the slight excess of chromate of potash. 

 The presence of soluble salts in the pressed precipitate, and its decom- 

 position by washing, determined us to the third method of analysis above 

 described (p. 197). Omitting the subsidiary weighings, we give the 

 essential figures of our first analysis of this substance. 



Weight of the precipitate analyzed 0.5997 gram. 



" " water found 0.1948 " 



« " oxygen " 0.0197 « 



" « soluble salts found 0.1969 « 



« " mixed precipitate of CrgOa + AlA 0.1883 " 



It will be remembered that the water and the precipitate of chromic 

 oxide and alumina are actually weighed, the oxygen being determined 

 by subtracting the water found from the total loss by ignition, and the 

 soluble salts by subtracting the weight of the precipitate of chromic 

 oxide and alumina from the weight of the whole residue after ignition. 

 The amount of chromate of alumina corresponding to the oxygen lost 

 by ignition may be calculated as follows : — ignition drives off half of 

 the oxygen originally combined with the chromium ; hence 



Oxygen in the chromic acid of precipitate = 0.0394 



Chromic acid corresponding to this oxygen = 0.0825 



Chromate of alumina " " chromic acid = 0.1666 



This calculation rests entirely upon the single determination of the 

 oxygen, and an inspection of the steps of the j^rocess will show that any 

 error in the amount of oxygen, or rather in the water determination on 

 .which the weight of oxygen depends, is multiplied by eight and a frac- 

 tion in the calculated amount of chromate of alumina. A considerable 

 discrepancy is therefore to be expected between the amount of chro- 

 mate of alumina so calculated, and the amount actually found in the 



