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

OBSIDIAN.

Approaching nearer the cliff and climbing over the masses of rock which lie at its base we shall find our interest increasing as the great beauty and variety of the lava are disclosed. At the southern end, where the road passes beneath the columns, the greater part of the rock is black, lustrous glass, or obsidian, so named after Obsidius or Opsius, its discoverer, in Ethiopia, according to Pliny, who says that when inlaid in chamber walls in the form of mirrors it reflects shadows instead of images. He also states that the ancients made signet stones of it.¹

---

¹C. W. King: The Natural History of Gems, London, 1867, p. 209.

---

King, in the work just cited, calls attention to the fact that the Peruvians also used the same stone for mirrors.

When broken it flies into sharp, angular fragments, with razor-like edges which are quite transparent and colorless in the thinnest places. It is this quality of the stone, together with its hardness, very nearly that of quartz, which makes it a favorite material for use as knives, spearheads, and arrowpoints by semi-civilized people. It was in common use among the Aztecs at the time of the conquest of Mexico by the Spaniards, as we learn from Prescott,² who, in describing the various implements used by these people, says:

They employed another tool, made of itztli, or obsidian, a dark, transparent mineral, exceedingly hard, found in abundance in their hills. They made it into knives, razors, and their serrated swords. It took a keen edge, though soon blunted. With this they wrought the various stones and alabasters employed in the construction of their public works and principal dwellings (p. 143).

Their weapons were slings, bows and arrows, javelins, and darts. * * * These various weapons were pointed with bone or the mineral itztli (obsidian), the hard, vitreous substance already noticed as capable of taking an edge like a razor, though easily blunted. * * * Instead of a sword they bore a two-handed staff, about three feet and a half long, in which, at regular distances, were inserted, transversely, sharp blades of itztli — a formidable weapon, which an eye-witness assures us he had seen fell a horse at a blow (p. 433).

---

²William H. Prescott: History of the Conquest of Mexico, vol. 1, Phila., 1874.

---

Another character of obsidian, which it shares with other homogeneous glasses, is the shelly or conchoidal fracture which is developed when it is broken by the blow of a hammer, the undulations of the surface of fracture spreading in arcs of concentric circles around the point of concussion. Occasionally the fracture takes the form of a cone, whose apex is at the point struck. Still more rarely a natural spherical sundering is observed, where thin layers like those of an onion encase a solid nucleus.

The color of this obsidian for the most part is jet black, but much of it is mottled and streaked with bright brownish red and various shades of brown, from dark to light yellowish brown, purplish brown, and olive green. The brilliant luster of the rock and the strong contrast of these colors with the black are very striking. In places the glass has been broken into small, angular pieces and cemented together, producing a many-colored and beautiful breccia. Some of the obsidian shows a fine, satin luster in certain positions. This is produced by the reflection of light from the walls of long, slender gas cavities which fill that portion of the glass that approaches the surface pumice. Another luster is occasionally exhibited in deeper-seated obsidian; it is a golden sheen which under a lens resolves itself into thin beams of red and yellow light, apparently reflected from minute cracks along the surface of microscopic shreds of brown and yellow glass.

SPHERULITES.

Through the black and red glass are scattered dull, bluish-gray patches and bands and round, gray and pink masses, looking like concretions, some of them being hollow or porous. The effect of these spherulitic forms is still further to vary the appearance and beauty of the rock and to make it the most conspicuous and characteristic variety of volcanic lava known. Similar spherulitic obsidian occurs in many parts of the world, notably in the Lipari Islands in the Mediterranean, in New Zealand, Mexico, and the western United States. The last-named occurrence will be more particularly noticed at the end of this paper. Owing to the great variety and the freshness of the spherulites and the perfection and beauty of the lithophysæ occurring at Obsidian Cliff, as well as to the interesting results derived from a careful study of them, they will be described with considerable detail, in order to give a clear idea of their structure and composition.

The simplest forms of the macroscopic spherulites—that is, those visible to the unaided eye—are small, dark-blue spherules about the size of a mustard seed, embedded in the black obsidian. When broken they appear lighter gray within, have a dense, porcelain-like texture, and show slight indications of a radially fibrous structure. They are mostly located along fine lines of minute punctures on the surface of the obsidian. The small, blue spherules are generally crowded together along these lines, or more properly along the planes of which these lines are the traces, the layers of spherules agreeing in all their bending and contortions with the fundamental planes of internal flow in the glass. A number of layers will sometimes lie close together, with the thinnest possible sheet of black glass between, or they will unite in a band a quarter of an inch or more thick, whose surface is covered with protruding hemispherules. Isolated clusters of blue spherules occur, making compound spheres, and more rarely these groups are prolonged in one direction, forming parallel ropes through the black glass.

The surfaces of the spherules are often brown or red and constitute planes of weakness between the spherules and inclosing glass, along which the two frequently separate with ease, leaving a dull, pitted surface on the obsidian. The arrangement of the spherules in the plane of flow is then seen to be quite irregular, though occasionally in arborescent figures.

Spherules about the size of peas have an agate-like banding in concentric shells, combined with a radially fibrous structure. Their form is more or less spherical, sometimes being depressed on one side and looking like miniature tomatoes or else prolonged into oblong gourd shapes. More frequently they are aggregated in botryoidal and kidney-shaped forms. Their surface, when relieved of the obsidian, has a delicate, velvety bloom like a peach, which in rich shades of brown and terra-cotta contrasts finely with their black, glassy matrix.

The larger spherules, an inch or more in diameter, are mostly lighter colored, in various shades of reddish gray, sometimes with a blue center. They appear of a more earthy texture than the small ones, but have a fine, radially fibrous structure, with a satin luster; in some cases there is a granular, spotted appearance toward the outer portion, and frequently a distinctly concentric structure is present, the shells being either broad and dense or of the most delicate thinness. The surfaces of these spherules are often ribbed with rings running parallel to the flow planes of the rock, closely resembling the surface of concretions in sedimentary rocks. Their shapes vary from spheroidal to flattened disks and hemispheres, and are also in irregular sectors and plume-like forms produced by the interference of spherules growing close together.

Through all these spherules, large and small, run delicate lines of banding parallel to and in continuation of the planes of flow inherent in the rock; sometimes they are only recognizable with the aid of a lens. This indicates that the spherules were developed in the glass in the same relative positions they now occupy; that is, they must have been formed after the lava came to rest. Frequently the largest spherules were formed earlier than the smaller ones, but this is by no means a fixed rule, for large, red hemispherules have often developed on a layer of small, blue ones, and discoidal ones between such layers.

Hollow spherulites.—Finally, the spherules are not all solid, the larger ones not at all so. Those looking earthy and densest have microscopic spaces between the fibers and grains composing them. Others have a fine granular appearance and glisten from the crystals of which they are made up. Most of them have porous or open cavities within their mass, the periphery often forming a solid shell or crust like the rind of a cantaloupe. These cavities ramify through the heart of the spherule and are coated with brilliant crystals. The porous spherules resemble pithy berries, while the central mass of the more open ones appears to have shrunken and cracked apart like the heart of an overripe watermelon. This is shown to a certain extent in Pl. XII, Figs. 1 and 5. The fibers and nodules of the mass are often very distinct and coarse and its color is nearly white. In many cases the cavity is confined to the limits of a single spherule, detached ones with perfectly solid exteriors being often hollow within. The isolated spherules in dense, black obsidian without the microscopical trace of cracking are sometimes half hollow, presenting the purest white skeleton of fibers and nodules, or consist of concentric, crystalline shells dotted with minute pellets. This form is represented in Pl. XII, Fig. 3. The shells are usually so delicate that parts of them are loosened and fall out when the obsidian is broken. On the white and beautifully frosted substance of these hollow spherulites rest the honey-yellow crystals of fayalite, as yet unattacked by atmospheric agencies. Cavities are less frequent in the small, blue spherules, though occasionally the smallest are white and porous, and along the spherulitic layers or through the black glass run crinkled bands of small cavities, or porous layers, with a white, gray, or pink coating.

PL. XII LITHOPHYSÆ.

LITHOIDITE.

Besides the thin, blue layers with spherulitic structure, occurring in the obsidian, there are light-gray ones of a more crystalline or porcelain-like nature. As these become more frequent the rock assumes a lithoidal or stony appearance and grades into purplish-gray rock, finely banded with blue, having occasional layers of black obsidian running through it. At the northern end of the cliff, as already noticed, there is very little black glass left and the rock is an excellent lithoidal rhyolite or lithoidite. This lithoidite is a light purplish-gray rock, which shows, on cross-fractures, delicate bands of light and dark colored layers. The former are crystalline, with small cavities scattered along them, which form planes of weakness and permit the rock to split into thin plates often one-sixteenth of an inch in thickness. The dark layers are microspherulitic and dense, showing in many cases a system of fine parallel cracks, all of them running in one direction and perpendicular to the plane of the layer. These cracks are at quite uniform distances apart in any one layer, usually half an inch and less, being closer together the thinner the layer. Upon examination it is seen that the direction of the cracks is at right angles to the streaks of color in the rock, which mark the direction of flow. From this position of the cracks, which are the result of the shrinkage of the rock upon cooling and which must lie at right angles to the direction of maximum strain, we see that along the line of flow at the time of consolidation the tension was greater than in other directions, which was probably due to a pulling stress exerted in that part of the lava sheet that stretched down the slope from the plateau above. In places the rock breaks into blocks which bear so striking a resemblance to silicified wood that one is easily deceived at a little distance.

LITHOPHYSÆ.

The lithoidal rock is as full of spherulitic forms as the obsidian, but it appears more porous and contains a multitude of hollow spherules of the utmost delicacy and beauty. An idea of their great abundance is given by Pl. XIII, Fig. 2, which was drawn from a slab of lithoidite and is the natural size. Most of them are hemispherical and consist of a group of concentric shells which curve one over another like the petals of a rose (Pl. XII, Fig. 4, and Pl. XIII, Fig. 1). The shallower ones present small, rose-like centers surrounded by thin, circular shells (Pl. XIII, Fig. 2). The disks are sometimes oval and sometimes composed of several sets of shells which have started from centers near together and developed in sectors, giving a scalloped form to the curves (Pl. XIII, Fig. 2). Others are eccentric or send out long, curving arms, cross-walled like a chambered ammonite (Pl. XII, Fig. 2).

PL. XIII LITHOPHYSAE IN THE LITHOIDITE OF OBSIDIAN CLIFF, NATURAL SIZE.

The partition walls are generally very thin and often close together, in one instance fifty occurring within a radius of two inches. They are very fragile and crumble under the touch, being made up of small and slightly adhering crystals with brilliant, glistening faces.

A fine example of a lithophysa is shown on a natural scale in Pl. XIII, Fig. 1. The rose-like center is surrounded by delicate shells. Those outer portions to the right and left which still remain are somewhat massive, though finely porous and crystalline, and are traversed by well marked shrinkage cracks. The contraction of the massive portions is clearly indicated by these cracks, which gape open from the base to which the substance of the lithophysa adhered. The largest forms are a foot or more in diameter and are very suggestive of wasps' nests.

These hollow structures have been called Lithophysen by von Richthofen,¹ because the rock appears to have been inflated or expanded like bubbles. He described those found in the rhyolites of Hungary and considered that the viscous rock had been expanded by steam. A review of his observations on their nature and origin will be given in its proper connection.

---

¹Studien aus den ungarisch-siebenbürgischen Trachytgebirgen: Jahrbuch k. k. geol. Reichsanst., vol. 11, 1860, p. 180.

---

But these lithophysæ cannot be considered as having been actually expanded by gas or steam, however much they resemble the bursting bubbles which rise to the top of boiling mud, a sight familiar to all who visit the Yellowstone region. They bear an intimate relation to spherulites, as the stone spherules are called, which becomes evident upon comparing the various structures exhibited by each.

Corresponding to the complete spherulite with simple, radially fibrous structure and no concentric banding, there are similar spherulites partly hollow at the center or irregularly so through their mass, as in Pl. XIV, Fig. 1. The simple spherulite which is traversed by parallel bands, in continuation of the planes of lamination in the rock, has a corresponding porous form, in which these bands become partition walls between the porous or hollow portions (Pl. XIV, Fig. 2). Spherulites with concentric banded structure have more or less hollow varieties where the denser bands remain as thin shells coated with crystals (Pl. XIV, Fig. 3), and among these are some crossed by the parallel bands of flow structure, which are clearly recognized in the hollow forms (Pl. XIV, Fig. 4). The hemispherulite with concentric bands, when developed in a narrow layer of glass, spreads out in the shape of a flattened disk, the bands farthest from the center being only short arcs of circles. The hollow form corresponding to this is the most characteristic lithophysa found in the lithoidite of Obsidian Cliff, which has already been described and is shown in cross-section in Pl. XIV, Fig. 5. Lithophysæ intersected by the bands of lamination of the matrix also occur (Pl. XIV, Fig. 6), the two systems of delicate walls being distinctly independent of each other. Finally, partial spherulites in the shape of sectors and plume-like forms are represented among the irregularly developed lithophysæ.

PL. XIV VARIETIES OF LITHOPHYSÆ.

From the foregoing it is evident that lithophysæ differ from spherulites neither in outward form nor general structure, but in the nature and continuity of the material composing them. At first sight it would appear as though the substance of the spherulites had been attacked by some corrosive agent which had partially reduced it and deposited in the resulting cavities the crystals already alluded to. How far this may have been the case will appear in the sequel.

MINERALS COMPOSING LITHOPHYSÆ.

The minerals which coat the walls of the lithophysæ and make up the material which is found in the perfectly fresh ones are quartz, tridymite, feldspar, fayalite, and magnetite. The size of the crystals bears a direct relation to the size of the cavity in which they have been developed when the lithophysa is isolated in a glassy matrix. The crystals are very minute in small lithophysæ and larger in the large ones, the character of the minerals remaining the same. Where several lithophysæ are connected or where a system of cavities traverses the rock a separation of the minerals takes place, certain varieties being deposited in particular localities, producing larger crystals whose size bears no relation to that of the cavity in which they occur.

Quartz is in prismatic crystals characteristic of this mineral when freely developed in cavities. The crystals are often doubly terminated with two sets of rhombohedrons, the ordinary set, ± R, and a steeper one, and occasionally a scalenohedron. The prism faces are strongly striated transversely. The substance of the quartz is very pure and transparent, with numerous gas cavities inclosed in it.

Tridymite, the second form of silica found in lithophysæ, is in delicate, six-sided plates. The larger crystals, which maybe recognized by the unaided eye, are 0.5mm (one-fiftieth of an inch) broad and have rather stout, tabular forms with a very simple combination of faces. They are easily confounded with the feldspars, on account of their form as well as of their substance, which is transparent and colorless, with scattered gas inclusions. Between crossed nicols they show the optical anomalies characteristic of tridymite. The crystals are frequently single or grouped in clusters, which latter, when composed of microscopic individuals, appear as minute pellets. The tridymite is deposited along with prismatic quartz in varying proportions, sometimes one occurring to the exclusion of the other.

Two forms of feldspar have so far been recognized. One is in almost microscopic crystals, which make up the coarser fibers in the lithophysæ. They are short and stout, with the form characteristic of adular, that is, in rhombic prisms bounded by the unit prism faces and terminated by a striated, compound face formed of the basal plane and an orthodome. These short prisms are attached to one another end to end, making crooked and forking branches of feldspar.

Another form is found in the lithophysæ of the lithoidal rock and the cavities connected with them, where the largest crystals occur. They are thin, tabular crystals about 1mm broad, flattened in the plane of the base, and bounded by the clinopinacoid, prism, and two orthodomes. Such a development of the basal plane is quite uncommon for orthoclase, of which species these feldspars prove to be a somewhat doubtful variety. Mr. S. L. Penfield, of the Sheffield Scientific School, has measured and figured the crystals, besides making a chemical analysis of them.

The simple and usual form is represented in Fig. 51, though many of the crystals are much thinner than this. They are frequently twinned according to the Manebacher law, as in Fig. 52, and very rarely according to the Baveno law.

FIG. 51. Crystal of soda orthoclase.

FIG. 52. Twinned crystal of soda orthoclase.

The axial ratio of the crystals was found from the following measurements:

Additional angles measured and calculated were as follows:

Angle c ^ b could not be measured very accurately, owing to poor reflections from the faces. The reflections from x also were very faint. These measurements show that the crystals belong to the monoclinic system, and their habit is that of orthoclase feldspar, but their optical character is not that of a monosymmetric mineral. The thin, tabular crystals furnish plates parallel to the basal plane and readily permit of optical investigation in this position, which is a fundamental one for a monosymmetric crystal. The edge made by the base and chinopinacoid is sharply defined, and the edge between the orthodome and base is seen to be exactly at right angles to it, as closely as can be measured by the cross-wires of the microscope. But between crossed nicols these basal plates do not extinguish the light parallel to the trace of the clinopinacoid, as they should do if the crystals were monosymmetric, and in convergent polarized light it is found that the plane of the optic axes makes a small angle with the clinopinacoid, which in the first eleven feldspars examined ranged from 1° to 4°. In fifteen more feldspars examined with the wide-angle, immersion, Bertrand lens of Nachetet Fils, the inclination ranged from 1° 30' to 5° 12', in no case being 0°.

The inclination of the bisectrix to the edge c b measured on the clinopinacoidal face of cleavage pieces varied from +6° to +10°.

Optically the crystals are asymmetric and approach the recently described anorthoclase of Klein¹ and Forstner.² It is to be hoped that better material may be found which will furnish larger crystals for a thorough determination of the optical properties of these abnormal feldspars.

---

¹Ueber die Feldspath im Basalt von Hohem Hagen etc.: Göttinger Nachrichten, 1878, No. 14; Neues Jahrbuch für Mineral., 1879, p. 518.

²Ueber die Feldspäthe von Pantelleria: Zeitschr. für Kryst., 1883, vol. 8, p. 125.

---

The substance of the feldspars is perfectly transparent, with minute gas inclusions. The crystals exhibit a brilliant, blue opalescence that appears to be reflected from planes of microscopic parting parallel to an orthodome, indicated by delicate lines, which, on the basal plane, lie parallel to the axis b, and, on the clinopinacoid, make an angle of about 71° 42' with the axis a. This corresponds closely to the observations of Mr. Whitman Cross¹ on the sanidines in the rhyolite of Chalk Mountain, Colorado, and still further confirms the deductions of E. Reusch,² who ascribed the iridescent colors noticed in many labradorites and other feldspars to the effect of thin plates due to delicate parting in certain directions through the mineral.

---

¹Contributions to the Mineralogy of the Rocky Mountains, Whitman Cross and W. F. Hillebrand: Bull. U. S. Geol. Survey No. 20, 1885, p. 75.

²Ueber das Schillern gewisser Krystalle: Poggendorff, Annalen, vol. 116, p. 392; ibid., vol. 118, p. 256; ibid., vol. 120, p. 95.

---

The chemical composition of these feldspars is shown by the following analysis made on less than a gram of material, which was obtained by means of Thoulet's solution of iodide of potassium and iodide of mercury—a solution which may be prepared with a specific gravity of about 3, when it will float many of the rock-making minerals, the heavier ones falling to the bottom. By gradually diluting the solution with water its density may be lowered to that of the suspended minerals, which one by one will settle to the bottom, as the specific gravity of each is passed. In this way the feldspars were separated from portions of the rock mixed with them and fell between the specific gravities of 2.589 and 2.541.

Partial analysis, I, was made on 0.6002 gram, and, II, on 0.3581 gram.

The ferric oxide was derived from a little yellow oxide of iron which coated some of the crystals and may be omitted in discussing the analysis.

The ratio

indicates a slight excess of silica, probably due to a small amount of quartz or tridymite which adhered to the feldspar crystals. Eliminating this excess of silica, which, when calculated to bring the ratio of SiO₂ : Al₂O₃ : R₂O = 6 : 1 : 1, would be 5.25 per cent., the analysis becomes

which will reduce to

From the ratio of the alkalis in this analysis, the feldspar may be considered as an isomorphic mixture of orthoclase and albite in nearly equal proportions; its formula will then be Or₂Ab₁. It is the middle member of the orthochase-albite series, with apparently the crystallographic characters of orthoclase, though somewhat exceptionally developed, but with the optical characters of a triclinic feldspar—an anomalous combination of characters which requires further investigation.

Fayalite is in small, tabular crystals scattered through the lithophysæ, usually projecting from the walls of the cavities, its crystals being larger than those of the associated minerals. In most instances they appear as opaque, black crystals about 2mm and less in length; their form is flattened, or tabular, with orthorhombic symmetry. They frequently have a metallic luster, occasionally with brilliant, iridescent colors, mostly reds. Examination shows that these crystals are coated with ferric oxide, the interior of the crystals being transparent and of a light-yellow color. Perfectly fresh, unaltered crystals are found in the small lithophysæ isolated in the obsidian, where they have been preserved from the action of the atmosphere. The best crystals of this sort were found in compact, black obsidian, about half a mile north of Lake of the Woods, near the southern end of this obsidian sheet. They are in thin, square or rectangular plates, of a light honey-yellow color, perfectly transparent and free from inclusions of other minerals, but occasionally containing gas cavities. They show very slight pleochroism, pale greenish yellow parallel to the b axis and golden yellow parallel to the c axis. The cleavage parallel to the brachypinacoid is good, but a second at right angles to the first is less distinct and is probably in the plane of the macropinacoid, as in olivine.

Mr. S. L. Penfield has kindly determined and figured the crystallographic forms presented by the rectangular, tabular crystals from the locality north of the Lake of the Woods and the more elongated and pointed crystals from Obsidian Cliff. The measurements were made on a thin, tabular crystal 0.1mm thick and 0.8mm broad, which was broken across the prismatic zone. The observed forums were a (100, i-i), b (010, i-i), s (120, i-2), e (111, 1), d (101, 1-i), k (021, 2-i). Arrangement of planes is quite constant, as in Fig. 53. For fundamental angles the best two reflections were chosen:

The angles measured and calculated were:

FIG. 53. Crystal of fayalite.

The plane of the optic axes is parallel to the base, one of the bisectrices being normal to the macropinacoid, as shown by a polarizing microscope. Owing to the minuteness of the crystal examined, the divergence of the optic axes was not determined.

The opaque crystals from Obsidian Cliff show the same forms, with the additional basal plane, c, but are mostly developed as in Fig. 54. With the exception of the macropinacoid, the faces were too dull to give good reflections and the forms were identified by approximate measurements only.

FIG. 54. Crystal of fayalite.

This mineral is the same as that described in 1827 by Gustav Rose,¹ who measured the small crystals found in the lithophysæ in obsidian from Cerro de las Navajas, Mexico, which von Humboldt had collected.

---

¹Ueber den sogenannten krystallisirten Obsidian: Poggendorff, Annalen, vol. 10, 1827, pp. 323-326.

---

"The crystals," he says, "are very small, the largest half a line long and broad, and very thin, of a greenish and reddish-yellow color, transparent, but with a strong, vitreous luster." The forms of the crystals figured by him are exactly the same as those found at Obsidian Cliff, and the angles measured are almost identical. They are as follows:

From the similarity in crystallographic form and in manner of occurrence, it is probable that the crystals which Rose determined as olivine belong to the purely ferruginous variety, fayalite.

A chemical analysis of the coated crystals from Obsidian Cliff was made by Prof. F. A. Gooch, at that time in the chemical laboratory of the U. S. Geological Survey. All the material available was 0.24 gram.

Under the microscope the crystals were seen to carry a small amount of adhering quartz and to be coated with iron oxide. They were readily decomposed in hot hydrochloric acid with the separation of silica and yielded the following results:

Considering the ferric oxide as the opaque coating of alteration and the insoluble silica as the adhering quartz, the composition of the unaltered mineral will be:

which is essentially the composition of the unisilicate fayalite.

In the case of the unaltered crystals from near Lake of the Woods, a very careful qualitative test by Mr. Penfield showed that these crystals were an iron silicate containing no magnesia.

The magnetite in the lithophysæ is in microscopic grains which under the microscope appear to be octahedral crystals.