of which we have just spoken, were discovered by M. Nordenskiöld. The discussions upon their origin, which we hesitated at first to regard as terrestrial, sufficed to bring out the close analogies between them and the meteorites. The study of the latter bodies has, then, permitted us to penetrate by induction into the internal constitution of our globe, as if by a side-look into depths wholly inaccessible to direct observation. The last demarkation has thus been effaced, and a most intimate connection has been established between the masses thrown up from the interior of our planet and the celestial masses of which the meteorites bring us the fragments.
The analogies which we have pointed out between meteorites and the profound regions of our globe testify to the identity of the chemical actions, even in the formation of stars very distant from each other. In fact, a mineral generally suggests, in a precise manner, the circumstances under which it originated. We might say that in itself it tells the story of its origin, especially when it can be reproduced experimentally. We thus perceive how reason, assisted by experiment, can give us clews to the formation of the stars of which we possess fragments. Silica or silicic acid is a chemical agent, the energy of which becomes very considerable at high temperatures; it is also the characteristic element of numerous products formed in industrial furnaces, like glass, scoriæ, and slags, and of the lavas of volcanoes. All the silicates, artificial and natural, when free from water, or in the anhydrous state, denote the dominance of a high temperature over their formation.
Suppose that silicon and the metals were not originally combined with oxygen as they are now, either because the different elements were not near enough together in the primordial chaos, or their temperature was not high enough to permit them to enter into combination. When oxygen comes into action, it unites at first with the elements for which it has a predominant affinity, primarily silicon and magnesium, then iron and nickel; and, if the gas is not in excess, it leaves a residue composed of the less oxidable bodies. Iron and nickel would in that case be left in a free state, disseminated among the stony silicates. This is exactly what is observed in the meteorites; and it is also a fact which I have confirmed by experiment. By producing the conditions that have just been mentioned, I obtained an imitation in essential points of meteorites of the common type, with the production of a silicate of magnesia and protoxide of iron, having exactly the constitution of peridote.
Furthermore, one of the best-known every-day metallurgical operations, the decarbonization of cast-iron, or its transformation into malleable iron or steel, gives an analogous reaction and ends in a result of the same kind. Whether the process be carried on in little charcoal-furnaces, as in antiquity, or in puddling-furnaces, or, as by the Bessemer process, without the addition of any combustible, it is always the
