216 PROCEEDINGS OF THE AMERICAN ACADEMY 



analysis, the sum of the oxygen and of the mixed precipitate of chromic 

 oxide and alumina. That sum is in this analysis 0.208 gram. ; sub- 

 tracting the calculated amount 0.1 6G6 gram., we find a discrepancy of 

 0.0414 gram. By subtracting 0.005 gram, only from the weight of the 

 water determined, this error would be corrected, and the precipitate would 

 coincide exactly with the formula AI2O3 CrOs, or, as the figures stand, 

 its composition is very nearly that represented by the formula (Al203)3 

 (Cr03)2. Of course a method of analysis involving such a multiplica- 

 tion of an error, however small, is objectionable, but we were at a loss 

 to devise a better. A second analysis gave a similar result ; the dis- 

 crepancy between the calculated and the actual numbers was four 

 milligrammes larger than in the first analysis, and the figures of the 

 analysis gave a chromate of alumina a little more basic than that 

 represented by the formula (Al203)3 (Cr03)2. We attach very little im- 

 portance to any such formula. The fact seems to be that the normal 

 composition of the chromate of alumina precipitated by the double de- 

 composition of alum and chromate of potash is represented by the for- 

 mula AI2O3 CrOa, but that' the compound is a feeble one, and parts read- 

 ily with some of its chromic acid to the water of the solutions from 

 which it is precipitated ; hence the basic character of the precipitate 

 analyzed. How Fairrie obtained a thoroughly washed chromate of 

 alumina having the formula AI2O3 CrOa is a mystery upon which our 

 experiments have thrown no light. 



2. Chromate of Iron. We have already cited the observations of 

 Thomson concerning this chromate, which he found after washing to be 

 very basic, and also the statement of Maus concerning the acid chromate 

 of iron with four equivalents of chromic acid, obtained by digesting hy- 

 drated sesquioxide of iron in aqueous chromic acid. It remained for 

 us to analyze, as exactly as the circumstances permit, the unwashed 

 cln'omate of iron which is precipitated when chromate of potash is 

 mixed with perchloride of iron. The method of analysis was the same 

 as that used for chromate of alumina, with the disadvantage that the 

 error in the amount of oxygen found is multiplied by ten instead of by 

 eight in calculating the chromate of iron, because the atomic weight of 

 the sesquioxide of iron is larger than that of alumina. The stability 

 of the clu'omate of iron is greater, however, than that of the chromate 

 of alumina, and the i-esults of the analyses are very much nearer to the 

 formula Fe203 CrOs, than those of chromate of alumina to the formula 

 AI2O3 CrOa. The formula for the reaction whereby the brown chro- 



