Carbonates of alumina, glucina, iron, chromium, etc. 323 



bonate of soda, in very dilute solution, and dries the precipitate 

 over sulphuric acid. His results give the formula Cr^O,, CO^, 

 4H0. Barratt* proceeds in the same way as Wallace and ob- 

 tains the same result, which is also that obtained by Lefort. 



The substance which I analyzed was precipitated by carbon- 

 ate of soda, in slight excess, from crystallized chroriie-potash- 

 ahm, free from iron and alumina. In the first analysis, given 

 below, the substance was dried over sulphuric acid: in the others 

 not. The first analysis was made like those of carbonate of iron. 

 The wash-water from the oxyd of chromium, left in the bulbed 

 tube after ignition, was colored yellow from alkaline chromate. 

 This must have arisen from oxydation of a little oxyd of chro- 

 mium, by ignition in the air in contact with alkaline carbonate. 

 The solution was reduced by hydrochloric acid and alcohol, pre- 

 cipitated by ammonia, and the oxyd of chromium thus obtained 

 added to the rest. There still remained an error, however, from 

 the chromic acid having driven out carbonic acid frotn the car- 

 bonate of soda; and to°avoid this, the other determinations were 

 made in a stream of hydrogen. In this latter method there would 

 be a possible source of error, viz., a reduction of the alkaline 

 sulphate to sulphid. This, however, was extremely small, as the 

 Wash-water, on heating it with hydrochloric acid, gave off in 

 every case only a tra^e of sulphuretted hydrogen and either 

 remained clear, or became only very slightly milky, from pre- 

 cipitated sulphur. 



The analyses of a large number of preparations gave the fol- 

 lowing results : 



The composition of the first three precipitations comes pretty 

 near the formula, given above, while the other two vary lar-ely 

 f'-om it. These two were made in summer, while the others 

 fere made in cold weather, and it is possible that there may have 

 Jeen a greater loss of carbonic acid at the higher temperature. 

 I :im inclined to think, however, that this is not the case, as, m 

 the last preparation, the liquids were cooled, before precipita- 

 tion, and the precipitate left to press in a refrigerator, containmg 

 'J- It occurred to me that one cause of the varv.ng composition 

 Of the precipitates might be that basic compoauds might at tirst 

 go down, before the carbonate of soda was in exces-s as is the 

 ?ase, for example, when ammonia, not in excess, is added to so- 

 lutions of copper. To ascertain whether this was the case, an 

 aQalvsis was made of a precipitate, produced by pouring chrome 



