It appears from three analyses, that obsidian contains on an average 76 per cent of silica; from one analysis, that sphaerulites contain 79.12; from two, that marekanite contains 79.25; and from two other analyses, that pearlstone contains 75.62 of silica. (The foregoing analyses are taken from Beudant "Traite de Mineralogie" tome 2 page 113; and one analysis of obsidian from Phillips "Mineralogy.") Now, the constituent parts of trachyte, as far as they can be distinguished consist of feldspar, containing 65.21 of silica; or of albite, containing 69.09; of hornblende, containing 55.27 (These analyses are taken from Von Kobell "Grundzuge der Mineralogie" 1838.), and of oxide of iron: so that the foregoing glassy concretionary substances all contain a larger proportion of silica than that occurring in ordinary feldspathic or trachytic rocks. D'Aubuisson ("Traite de Geogn." tome 2 page 535.), also, has remarked on the large proportion of silica compared with alumina, in six analyses of obsidian and pearlstone given in Brongniart's "Mineralogy." Hence I conclude, that the foregoing concretions have been formed by a process of aggregation, strictly analogous to that which takes place in aqueous deposits, acting chiefly on the silica, but likewise on some of the other elements of the surrounding mass, and thus producing the different concretionary varieties. From the well-known effects of rapid cooling (This is seen in the manufacture of common glass, and in Gregory Watts's experiments on molten trap; also on the natural surfaces of lava- streams, and on the side-walls of dikes.) in giving glassiness of texture, it is probably necessary that the entire mass, in cases like that of Ascension, should have cooled at a certain rate; but considering the repeated and complicated alterations of nodules and thin layers of a glassy texture with other layers quite stony or crystalline, all within the space of a few feet or even inches, it is hardly possible that they could have cooled at different rates, and thus have acquired their different textures.
The natural sphaerulites in these rocks very closely resemble those produced in glass, when slowly cooled. (I do not know whether it is generally known, that bodies having exactly the same appearance as sphaerulites, sometimes occur in agates. Mr. Robert Brown showed me in an agate, formed within a cavity in a piece of silicified wood, some little specks, which were only just visible to the naked eye: these specks, when placed by him under a lens of high power, presented a beautiful appearance: they were perfectly circular, and consisted of the finest fibres of a brown colour, radiating with great exactness from a common centre. These little radiating stars are occasionally intersected, and portions are quite cut off by the fine, ribbon-like zones of colour in the agate. In the obsidian of Ascension, the halves of a sphaerulite often lie in different zones of colour, but they are not cut off by them, as in the agate.) In some fine specimens of partially devitrified glass, in the possession of Mr. Stokes, the sphaerulites are united into straight layers with even sides, parallel to each other, and to one of the outer surfaces, exactly as in the obsidian. These layers sometimes interbranch and form loops; but I did not see any case of actual intersection. They form the passage from the perfectly glassy portions, to those nearly homogeneous and stony, with only an obscure concretionary structure. In the same specimen, also, sphaerulites differing slightly in colour and in structure, occur embedded close together. Considering these facts, it is some confirmation of the view above given of the concretionary origin of the obsidian and natural sphaerulites, to find that M. Dartigues ("Journal de Physique" tome 59 1804 pages 10, 12.), in his curious paper on this subject, attributes the production of sphaerulites in glass, to the different ingredients obeying their own laws of attraction and becoming aggregated. He is led to believe that this takes place, from the difficulty in remelting sphaerulitic glass, without the whole be first thoroughly pounded and mixed together; and likewise from the fact, that the change takes place most readily in glass composed of many ingredients.