Geological Survey 7th Annual Report Obsidian Cliff, Yellowstone National Park

The probable origin of fayalite in so siliceous a rock as obsidian will appear from the following considerations:

MINERAL ASSOCIATION.

First. Its association with abundant quartz and tridymite and acid feldspars, with a small amount of magnetite, is not according to the laws which appear to govern the production of minerals by purely igneous fusion; for olivine is only found in igneous rocks low in silica and rich in iron and magnesia, and it is doubtful whether a purely iron olivine, like fayalite, has ever been found as an essential constituent of such rocks. At Fayal, where it has been found associated with trachyte, it would seem, from the description of Gmelin,¹ to occur as inclosed masses in the lava, for it is said to be full of bubbles in places, with the appearance of having been fused, and in the Mourne Mountains, Ireland, according to Delesse and Haughton, it occurs in drusy cavities in muscovite granite accompanied by beryl, chrysoberyl, fluorite, and topaz. The minerals accompanying olivine are usually lime-soda-feldspars, augite, and magnetite. Its occurrence in an obsidian with 75 per cent, of silica and less than 2 per cent of iron oxide is so contrary to common experience that the observations of Gustav Rose, in 1827, on an occurence of olivine in the obsidian of the Cerro de las Navajas, Mexico—which, as already pointed out, is similar to that at Obsidian Cliff—have been generally discredited by the most eminent petrographers.

---

¹Chemische Untersuchung des Fayalits: Poggendorff, Annalen, 1840, vol. 51, p. 160.

---

Secondly. The chief minerals which accompany the fayalite are those which have not been reproduced artificially under the ordinary conditions of purely igneous fusion. The experiments of Messrs. Fouqué and Michel-Lévy and others have demonstrated the failure of the purely igneous fusion of the chemical constituents of quartz and the acid felspars, orthoclase and albite, to produce crystals of these minerals similar to those found in acid rocks, though the more basic minerals have been so produced.

Thirdly. Quartz, tridymite, orthoclase, and albite have all been reproduced artificially by heating their component elements in the presence of superheated water in a closed tube: that is, by a form of aqueo-igneous fusion. The experiments bearing most directly on the case in hand are those of Messrs. Friedel and Sarasin,¹ who heated in a closed tube basic silicate of potassium and silicate of aluminium and obtained along with hydrous silicate of potassium crystals of orthoclase and quartz and, at higher temperatures, tridymite. The orthoclase crystals were in great variety of forms, some of which correspond to those accompanying fayalite in the lithophysæ. A similar treatment of silicates of sodium and aluminium gave rise to crystals of albite; but it is remarked that these were never accompanied by either quartz or tridymite and that a mixture of the alkalis produced orthoclase alongside of albite, although never in isomorphic combinations. Here, again, we see that the artificial methods of aqueo-igneous fusion, like those of purely igneous, or dry, fusion, have failed to reproduce intermediate varieties of feldspars corresponding to the isomorphic mixtures of distinct species, which are everywhere produced in nature, and of which the feldspar already described and figured from Obsidian Cliff is an example, and it seems probable that while all the conditions attending the natural processes of crystallization may not be comprehended in the artificial methods, still the far more gradual and less violent action of the same forces, which in particular instances in nature act feebly through long periods of time, may be an essential factor in the production of the infinite gradations in the composition of the minerals formed.

---

¹Bull. Soc. min´ralogie, 1879, vol. 2, p. 158; ibid., 1880, vol. 3, p. 171; Comptes rendus Acad. sci., Paris, 1881, vol. 92, p. 1374; ibid., 1883, vol. 97, pp. 290-294.

---

The experiments of K. von Crustschoff² show that under the influence of superheated water tridymite is produced at a higher temperature than quartz. Other experiments have given both quartz and tridymite in the same closed tube.

---

²Am. Chemist, 1883; Tschermaks mineral. Mittheil., vol. 4, p. 536.

---

Magnesian olivine has been produced³ by the action of steam and chloride of silicon on magnesium at a red heat under the pressure of the atmosphere, but as yet there seems to have been no attempt to reproduce iron olivine, which, however, occurs accidentally in many furnace slags. And finally magnetite⁴ has been reproduced by the action of aqueous vapor upon iron wire at high temperatures. From the foregoing we see that the minerals occurring in the lithophysæ at Obsidian Cliff are such as may be formed from a mixture of silicates under the influence of superheated water or steam.

---

³Stanislas Meunier: Comptes rendus Acad. sci., Paris, vol. 93, 1881, p. 737.

⁴De Haldat: Annales chimie, vol. 46, p. 70; Neues Jahrbuch für Mineral. 1833, p. e 680.

---

Fourthly. The experiments of Mr. Daubrée⁵ upon the effect of super-heated water on glass are of special interest in this connection. A sealed glass tube which contained only water was inclosed in an iron case also containing water, which was sealed and heated to redness for different lengths of time with differing results. Under certain conditions, part of the glass, a silicate of lime and soda, with 5 per cent, of alumina, was converted into a hydrous silicate, accompanied by a considerable increase of volume; part was reduced to a white, opaque mass distinctly fibrous, with a delicate banding parallel to the surface of the glass tube and resembling agate.

---

⁵Études synthétiques etc. Paris, 1879, pp. 154-179.

---

The surface of the tube in places was warped, blistered, and excoriated and frequently full of cracks, besides which a delicate foliation parallel to the surface of the glass was developed. In some instances the whole mass was reduced to powder. Examined under the microscope the altered glass showed minute nearly opaque spherulites, acicular microlites, small crystals and grains of pyroxene (diopside), and larger spherulites of a material like chalcedony. The surface of the glass was covered with prismatic crystals of quartz.

The correspondence of many of these characters with those observed in the obsidian of Obsidian Cliff is very striking and suggestive. The fibrous and delicately banded structure is similar to that of the smaller spherulites in the obsidian; the thin foliation, the gaping cracks, and encrusting quartzes are features characteristic of the lithophysæ. The microscopic forms of crystallization, though of different mineral nature, bear a great similarity, considering the difference in the chemical composition of the two glasses. Mr. Daubree calls attention to the large amount of alteration produced in the glass by a small amount of water, scarcely one-third of the weight of the glass.

The question then arises whether water was present in the molten obsidian during the crystallization of the spherulites, and, if so, to what extent.

As already pointed out, the larger spherulites composed of rays of feldspar and a cement of tridymite are filled with gas cavities, easily recognized by their spherical shape and behavior toward light. The smaller spherulites are often clouded with minute particles, which it was suggested were partly gas cavities. That this is the case is shown by the following experiment: A fragment of a small blue spherulite was heated in the flame of an oxyhydrogen blow-pipe till it melted, when portions of it puffed up into a pumiceons glass by the expansion of the inclosed gas. The same experiment was repeated on the compact, black obsidian itself. When melted it also to the pressure of the atmosphere was imprisoned in the deeper portions of the obsidian, probably combined with the glass, since the microscope fails to detect it.

The chemical analysis of the black obsidian and that of the dark blue spherulites show almost identical losses on ignition, in the former 0.66 per cent, and in the latter 0.33 per cent. The greater part of this is water, which has been determined directly as such, and careful analyses of the disengaged gases from other similar obsidians, made by Boussingault and Damour,¹ have shown its widespread occurrence, frequently accompanied by chlorine. It is to be noticed that a small amount of sulphur is present in the rock of Obsidian Cliff.

---

¹Annales chimie, Paris, 4th series, vol. 29, 1873, p. 547.

---

That the vapors existing in this obsidian were absorbed by the magma before its eruption is rendered highly probable by the experiments of Mr. Daubrée on the penetration of water through rock by capillary attraction against a counteracting pressure of steam and on the hydration of glass by superheated water. From the observations of Mr. Fouqué² at Santorin it is likely that at extremely high temperatures the elements of these vapors are dissociated.

---

²Santorin et ses éruptions, Paris, 1879, p. 232.

---

CHEMICAL EVIDENCE.

That the absorbed vapors in the rock were the sole agents affecting the crystallization of the denser spherulites and the production of the lithophysæ will appear from a consideration of the following chemical analyses, which were made of the rock of Obsidian Cliff for Mr. Arnold Hague, in charge of the Yellowstone National Park division of the U. S. Geological Survey. No. I is a black obsidian free from spherulites. No. II, red obsidian; No. III, small, dark-blue spherulities; and No. IV, white material forming small lithophysæ in solid, black obsidian.

Analyses I and II were made by Mr. Edward Whitfield, of the chemical laboratory of the U. S. Geological Survey, and III and IV were made by Mr. S. L. Penfield, of the Sheffield Scientific School.

Nos. I and II are almost identical, with less than 1 per cent. difference in silica. The chief point of interest lies in the relative amounts of ferric oxide and ferrous oxide present, the red obsidian having nearly all the iron in the form of ferric oxide, as already noticed.

Analyses III and IV are enough alike to be duplicates. They differ from those of the rock by being a little higher in silica and the alkalis and a little lower in the other bases. They are as close as could be expected, since the materials analyzed are from different parts of the lava flow. Other analyses might lessen this slight difference.

The main facts brought out by the analyses are that the chemical composition of the spherulite is essentially the same as that of the surrounding rock; it is then nothing more than a small portion of the magma which has crystallized with a particular structure; and, further, that the lithophysæ have the same composition as the dense spherulites, which shows that the transformation of a spherulite to a lithophysa can only be a modification of its structure, a rearrangement of its minerals, without any chemical addition or loss.

This rearrangement took place through agencies within the limits of the spherule affected and disconnected from other possible sources, for perfect lithophysæ occur isolated and hermetically sealed in dense, black obsidian. But where many such spherulites adjoined one another the cavities formed sometimes connected and spread irregularly, permitting the segregation of particular minerals in different places, which has been the case in the laminated, lithoidal portion of the lava flow.

Moreover, the hollow lithophysæ were formed prior to the consolidation of the surrounding matrix, for they are occasionally found with the outer shell crushed and the viscous matrix forced part way in, showing that the mass was at a very high temperature and still viscous when the modification of the spherulites took place.

CONCLUSION.

We may fairly conclude that the lithophysæ in the obsidian of Obsidian Cliff, with their contents of prismatic quartz, tridymite, adular-like and tabular soda-orthochase, magnetite, and well crystallized fayalite, are of aqueo-igneous origin and result from the action of absorbed vapors upon the molten glass from which they were liberated during the process of crystallization consequent upon cooling.

Apparent exceptions.—Cavities and lithophysæ occur, over which a distinct arching of the lamellae of the lithoidal rock is observed, the layers apparently accommodating themselves to the swelling of the hollows beneath; some appear simply as cavities coated with crystals of quartz, feldspar, and associated minerals; others have the concentric chambered structure of lithophysæ (Pl. XIV, Fig. 7). The same phenomena have been described by Prof. F. Zirkel in the lithoidite from the southeast shore of Taupo Lake, New Zealand. The hollow spaces in that rock are intersected by partition walls coated with crystals of quartz, feldspar, hornblende, and mica and in many ways correspond to the lithophysæ of Obsidian Cliff. At first sight it would seem that the expansion of a bubble of gas within the lava had occasioned the distention or displacement of its layers, but a careful study of portions of the rock which exhibit great contortion and plication of the layers makes it evident that in these cases the hollows occur beneath arches in the folds where there has been a local relief or diminution of pressure, which might allow the absorbed vapors to disengage themselves and bring about the conditions which produce hollow lithophysæ in connection with spherulitic development. In other words, the arching of the layers appears to have been the cause of the liberation of gases and the production of the cavity beneath, and not the result of expanding gases.

That such local inequalities of the layers could have existed within the plastic mass is evident from the fact that the moving lava was so viscous and stiff before its final halting that, as already noticed, in one place where the layers of flow reached a vertical position they pulled apart and solidified with gaping crevices between, the surfaces of the separate slabs consisting of stony fragments on a ropy, corrugated glass.