hours of the night more shooting stars pass by before midnight invisible. We must still long and patiently collect observations.

The principal characters of the solid masses which fall from the air, I believe I have treated of with tolerable completeness (Cosmos, vol. 1. p. 117.) in reference to their chemical relations and the granular structure, especially investigated by Gustav Rose in accordance with the state of our knowledge in the year 1845. The successive labours of Howard, Klaproth, Thénard, Vauquelin, Proust, Berzelius, Stromeyer, Laugier, Dufresnoy, Gustav and Heinrich Rose, Boussingault, Rammelsberg, and Shepard, have afforded a rich material, and yet two-thirds of the fallen meteoric stones, which lie at the bottom of the sea, escape our observation. Although it is striking that under all zones, at points most distant from each other, the aërolites have a certain physiognomic resemblance-in Greenland, Mexico and South America, in Europe, Siberia and Hindostan-still upon a closer investigation they present very great differences. Many contain of iron, others (Siena) scarcely 18; nearly all have a thin black brilliant and at the same time veined coating: in one (Chantonnay) this crust was entirely wanting. The specific gravity of some meteoric stones amounts to as much as 4.28, while the carbonaceous stone of Alais, consisting of crumbling lamelæ, showed a specific gravity of only 1.94. Some (Juvenas) have a doleritic structure, in which crystallized olivin, augite and anorthite are to be recognized separately; others (the masses of Pallas) afford merely iron, containing nickel and olivin; and others again (to judge from the proportions of the ingredients) are

27 The metals discovered in meteoric stones are, nickel by Howard, cobalt by Stromeyer, copper and chromium by Laugier, tin by Berzelius.

aggregates of hornblende and albite (Chateau-Renard), or of hornblende and labrador (Blansko and Chantonnay).

According to the general summary of results given by a sagacious chemist, Professor Rammelsberg, who has recently. occupied himself uninterruptedly, and as actively as successfully, with the analysis of aërolites and their composition from simple minerals, "the separation of the masses fallen from the air into meteoric iron and meteoric stones is not to be admitted in its strictest sense. Meteoric iron is sometimes found, though seldom, with silicates intermixed (the Siberian mass weighed again by Heis of 1270 Russian pounds, with grains of olivin), and on the other hand many meteoric stones contain metallic iron.

"A. The meteoric iron, whose fall it has been possible to observe only a few times (Hradschrina, near Agram, on the 26th of May, 1751, Braunau, 14th of July, 1847), while most analogous masses have already laid long upon the surface of the earth, possesses in general very similar physical and chemical properties. It almost always contains sulphuret of iron mixed with it in finer or coarser particles, which, however, do not appear to be either iron pyrites or magnetic pyrites, but a sulphuret of iron." The principal mass of such a meteoric iron is also not pure metal, but consists of an alloy of iron and nickel, so that this constant presence of nickel (on the average 10 per cent. sometimes rather more, sometimes rather less) serves justly as an especial criterion for the meteoric nature of the whole mass. It is only an alloy of two isomor phous metals, not a combination in definite proportions. There are also present in minute quantity: cobalt, manganese, magnesium, copper, and carbon. The last-mentioned substance is partly mixed mechanically, as difficultly combustible graphite;

28 Rammelsberg, in Poggendorff, Annalen, vol. lxxiv. 1849,

p. 442.

partly in chemical combination with iron, and therefore analogous to many kinds of bar-iron. The principal mass of the meteoric iron contains also always a peculiar combination of phosphorus with iron and nickel, which on the solution of the iron in hydrochloric acid, remains in the form of silver-white, microscopic, crystalline needles and laminæ.

"B. The meteoric stones, properly so called, it is customary to divide into two classes, according to their external appearance. The stones of one class present, in an apparently homogeneous mass, grains and splinters of meteoric iron, which are attracted by the magnet, and possess entirely the nature of that found in larger masses. To this class belong, for example, the stones of Blansko, Lissa, Aigle, Ensisheim, Chantonnay, Klein-Wenden near Nordhausen, Erxleben, Château-Renard, and Utrecht. The stones of the other class are free from metallic admixtures, and present rather a crystalline mixture of different mineral substances; as, for example, the stones of Juvenas, Lontalax, and Stannern.

"Since the time that Howard, Klaproth, and Vauquelin first instituted the chemical investigation of meteoric stones, for a long time no regard was paid to the fact that they might be mixtures of separate combinations; but they were examined only for their total constituents, and it was considered sufficient to draw out the iron by the magnet. After Mohs had directed attention to the analogy between some aërolites and certain telluric rocks, Nordenskjöld endeavoured to prove that the aërolite of Lontalax in Finland consisted of olivin, leucite, and magnetic iron ore; but the beautiful observations of Gustav Rose first placed it beyond doubt that the stone of Juvenas consists of magnetic pyrites, augite, and a feldspar, very much resembling labrador. Guided by this, Berzelius endeavoured, in a more extended essay (Kongl. Vetenskaps-Academiens Handlingar für 1834), to eliminate also by chemical methods the mineralogical nature of the separate combina

tions in the aërolites of Blansko, Chantonnay and Alais. The road happily pointed out by him beforehand has subsequently been abundantly followed.

66 a. The first and more numerous class of meteoric stones, those with metallic iron, contain this disseminated through them, sometimes in larger masses, which occasionally form a skeleton, and thus constitute the transition to those metecric masses of iron in which, as in the Siberian mass of Pallas, the other materials disappear more considerably. On account of the constant presence of olivin, they are rich in magnesia. The olivin is that part of the meteoric stone which is decomposed when it is treated with acids. Like the telluric, it is a silicate of magnesia and protoxide of iron. That part which is not attacked by acids is a mixture of feldspathic and augitic matter, whose nature admits of being determined solely by calculation from its total constituents, as labrador, hornblende, augite, or oligoclas.

"B. The second much rarer class of meteoric stones have been less examined. They contain partly magnetic iron ore, olivin, and some feldspathic and augitic matter; some of them consist merely of the two last mentioned simple minerals, and the feldspar tribe is then represented by anorthite.0 Chrome iron ore (oxide of chromium and protoxide of iron) is found in small quantity in all meteoric stones; phosphoric acid and titanic acid, which Rammelsberg discovered in the very remarkable stone of Juvenas, perhaps indicate apatite and titanite.

"Of the simple substances hitherto detected in the meteoric stones, there are 18:1 oxygen, sulphur, phosphorus, carbon,

30 Shepard, in Silliman's American Journal of Science and Arts, ser. ii. vol. ii. 1846, p. 377; Rammelsberg, in Poggend. Ann. Bd. lxxiii. 1848, p. 377.

31 Compare Cosmos, vol. i. p. 118.

silicium, aluminum, magnesium, calcium, potassium, sodium, iron, nickel, cobalt, chromium, manganesium, copper, tin, and titanium. The proximate constituents are: (a) metallic : nickel-iron, a combination of phosphorus with iron and nickel, sulphuret of iron and magnetic pyrites; (b) oxidized: magnetic iron ore and chrome iron ore; (c) silicates: olivin, anorthite, labrador and augite."

In order to concentrate the greatest number of important facts separated from hypothetic conjectures, it still remains for me to develop the manifold analogies which some meteoric stones present as rocks with older, so-called trap rocks, (dolerites, diorites, and melaphyren), with basalts and more recent lava. These analogies are so much the more striking, as "the metallic alloy of nickel and iron, which is constantly contained in certain meteoric masses," has not hitherto been discovered in telluric minerals. The same distinguished chemist whose friendly communications I have made use of in these last pages, enters fully into this subject in a special treatise," the results of which will be more appropriately discussed in the geological part of the Cosmos.

Zeitschrift der deutschen geolog. Gesellschaft, Bd. i. p. 232. All the matter in the text from p. 593 to p. 596, which is between inverted commas, was taken from the manuscript of Professor Rammelsberg (May, 1851).