Cavendish(1766) 155 cubic inches.
Priestley,from 147 to162
Lavoisier 163
Vandermonde, Berthollet, and Mongemax.176
Vauquelin160 to179
Dr. Thomson 163
My own Experiments give160
Mean164= 82 oxygen
= 27.9 grains.

By precipitating the oxide, and drying it, nearly the same result may be obtained, as 100 iron will yield 128 oxide. This oxide is magnetic.

2. Intermediate or red oxide. This oxide may be obtained in various ways. First by calcining the sulphate or nitrate of iron. Second by precipitation from old solutions of the salts of iron; the precipitate is yellow at first, being perhaps a hydrate; but when dried and heated it becomes brown-red. Third, by calcining iron or repeatedly exposing iron filings to a red heat, and trituration. Fourth, by treating a solution of the sulphate or other salt of the protoxide with oxymuriatic acid, or oxymuriate of lime till oxymuriatic acid is evolved; then precipitating the oxide which is thus converted into the red. Fifth, by agitating water containing the green oxide recently precipitated, with oxygen gas. The red oxide is not sensibly magnetic.

The quantity of oxygen in the red oxide may be investigated in various ways, and it is generally allowed that they all concur in giving 42 on 100 iron. The one which I have used peculiarly, and prefer both for ease and accuracy, is to find the quantity of oxymuriatic acid gas necessary to saturate a given portion of the green sulphate. I take for instance 100 measures of 1.149 green sulphate, which I know to contain 8 grains of black oxide; this I find absorbs nearly 13 hundred measures of oxymuriatic acid gas before the acid smell is developed; the oxygen corresponding to this quantity of acid is known to be near 660 measures, = .88 grain. (See Vol. 1, p. 308.) Hence, if 8 ∶ .88 ∷ 128 ∶ 14; or 128 black oxide acquire 14 or become 142 when converted into the red oxide. This fact being established, I find it very convenient to make use of the oxymuriate of lime instead of the acid gas, adopting the solution of green sulphate of iron as a test of the quantity of oxymuriatic acid in a given volume of any solution of oxymuriate of lime.

The quantity of oxygen in the red oxide of iron may be inferred, but not so satisfactorily, from the nitrous gas obtained during the solution of iron in nitric acid. In order to obtain the most gas from a given quantity of the materials, they should be so proportioned as to produce saturation nearly. If an excess of acid be used, it absorbs the nitrous gas in part; and if an excess of iron, it is not all dissolved. I took 50 grains of iron filings and 600 measures of 1.15 nitric acid; these were put together in a gas bottle and by the assistance of a little heat a quantity of nitrous gas was obtained equal to 12 grains in weight, allowing the sp. gr. of the gas to be 1.04 (air being 1); all the iron was dissolved except a few atoms, and the solution was slightly acid; the whole of the oxide was red when precipitated by lime water. Now 50 grains of iron take 21 of oxygen to form the red oxide, and these correspond to 24 of nitrous gas, which is just twice the quantity obtained; one half of the gas generated then remains in combination with the iron, even when the constituents of the salt are proportioned so as to produce mutual saturation. I was in expectation that the quantity of nitrous gas retained might be converted into nitric acid by oxymuriate of lime, and hence might be determined; but in this I was disappointed. When oxymuriate of lime is added to the liquid, a pungent gas is liberated, the nature of which I have not determined. Thinking it might in part be owing to the iron, I transferred the acid to soda, by decomposing the nitrate of iron by the carbonate of soda; this nitrate of soda however, when treated with oxymuriate of lime, exhibited the same phenomenon as the nitrate of iron. When an acid is added the oxymuriatic acid itself is given out. These results will require further consideration. At present I am inclined to think the pungent gas is one atom of nitrous and one of oxygen or what I formerly considered as nitric acid. (See Vol. 1, plate 4, fig. 27.)

Some authors have found as they conceive, other oxides of iron, containing less or more of oxygen than the above; thus Darso finds by calcination from 15 to 56 oxygen on 100, (Nicholson’s Journ. Vol. 17); but there is great reason to believe that uncertainties must exist in his mode of experimenting sufficient to account for the anomalies observed. This author has suggested some doubt whether the oxygenous gas naturally contained in water has any effect on the salts with green oxide of iron. I have ascertained that point by repeated experiments, and can assert that the oxygen in water immediately unites to the green oxide of iron to convert it into red, and that the green sulphate may be used as an accurate test of the quantity of oxygen in water. When pure green sulphate of iron is dropped into water and then the oxide precipitated by a gradual addition of lime water, it falls down yellow in proportion to the oxygen in the water, which may be increased 3 or 4 times by artificial impregnation. If the oxygen of the water be previously saturated with nitrous gas, then the oxide is wholly precipitated green.

Gay Lussac, in the 80th Vol. of the Annal. de Chimie, asserts that an oxide of iron containing 37.8 oxygen upon 100 iron is always obtained when iron is burned in oxygenous gas, and still more effectually when iron is oxydized by water or steam. If this oxide exist in the proportions stated, it must be a compound of 1 atom of the protoxide and 2 of the red oxide, which would give 37.3 oxygen on 100 of iron.

From the above facts and observations it is evident the atom of iron must be considered as weighing 25, (and not 50 as already given, Vol. 1, page 258); the protoxide is 32, and the intermediate or red oxide is 2 atoms protoxide and 1 of oxygen = 71.

11. Oxides of Nickel.

1. Protoxide. It appears to be ascertained from the experiments of Proust (Journ. de Physiq. 63—442), Richter (Nichols. Jour. 12.), Tupputi (An. de Chimie 78.), and Rolhoff (An. of Philos. 3—335.), that the protoxide of nickel consists of 100 metal and from 25 to 28 oxygen. My experiments on the solution of nickel in nitric acid give me 14 grains nitrous gas, corresponding to 12 oxygen, in the solution of 44 grains of nickel; this gives 100 nickel to 27 oxygen, which I adopt as agreeing with the mean of the beforementioned results. This oxide may be obtained by precipitation from a solution of nitrate of nickel; it is at first white, being then a hydrate; when dried in a moderate temperature it becomes yellowish; after this, being heated to a cherry red, it loses from 20 to 24 per cent. of water and becomes of an ash grey colour: this is the only oxide of nickel soluble in acids, and must therefore be deemed the protoxide: hence we have 27 ∶ 100 ∷ 7 ∶ 26, nearly, for the weight of an atom of nickel; and not 25 or 50, as estimated at page 258. Vol. 1.