218 PROCEEDINGS OF THE AMERICAN ACADEMY 



Chromate of iron found = FeaOj -|- CraOa -|- O weighed *== 0.3346 

 Excess of the calculated chromate of iron over the found = 0.0251 



This error would be corrected by a supposed loss of little more than 

 two milligrammes in the water determination. This second analysis, 

 therefore, confirms the view that the normal composition of this chx*o- 

 mate of iron is represented by the formula Fe^Oa CrOj. 



It is unnecessary to call attention to the great similarity of the prop- 

 erties of the chromates of chromic oxide, ferric oxide, and alumina ; 

 their normal composition is represented by the formula R2O3 CrOs, but 

 they are such feeble compounds as seldom to be found of the exact 

 constitution which the formula indicates.* Of the three, the chromate 

 of chromic oxide is the most stable, and we have been able to prepare 

 it with the exact theoretical composition, from which it so easily varies. 

 It should be observed, that a variation in the composition of these chro- 

 mates, amounting to even ten per cent of their chromic acid, does not 

 affect in the least their external properties, and it is by no means im- 

 possible that they are examples of that variation in the law of definite 

 proportions which has been ably discussed by Professor J. P. Cooke 

 in tlie Memoirs of this Academy ; t but unfortunately, as must often be 

 the case, the limits of error of our methods of analysis are wider than 

 those of the supposed variation, and moreover, on this border-land 

 between the kingdoms of Chemical and of INIechanical Force, there 

 must always be doubt and difference of oi)inion concerning the cause 

 to which a proved effect is to be ascribed. 



3. Chromate of 3Ianganese. The existence of this chromate has 

 been so clearly demonstrated by i)revious observers that we have 

 thought it unnecessary to add any experiments of our own. On mix- 

 ing a solution of sulphate of the protoxide of manganese with a solution 

 of chromate of potash, a chocolate-colored precipitate subsides after 

 some time, to which Warington J assigned the formula (MnO)2 CrOj 

 2 HO. Bensch § also analyzed a similar precipitate, for which he 

 gives the same formula. In Warington's analysis the amount of pro- 

 toxide of manganese is larger than the compound which his formula rep- 



* The monochromatc of glucina mentioned by Gmelin (Handbook, IV. 155, 

 Cavendish Soc. Ed.) evidently belongs to this class, 

 t New Series, V. pp. 348, 352 (1854). 

 X Phil. Mag., [3] XXI. 380 (1842). 

 § Pogg. Ann., LV. 98 (1842). 



