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



Obsidian Cliff is especially remarkable for the development of prismatic columns, which form the southern end of the mass. Such columns are of common occurrence in basalt the world over, the most familiar localities being the Giant's Causeway and Fingal's Cave, on the coast of Ireland. There are others in central France and Germany and throughout the great volcanic area in western North America, especially along the Columbia River and in the Yellowstone Park; they are in almost every basalt flow in fact, so that columnar structure is often erroneously termed basaltic structure. But columnar structure is by no means confined to such rocks, being found in all varieties, from the basic to the acidic, though less frequently in the latter. Columnar structure is well developed in the rhyolites of the Yellowstone Park. Instances of its occurrence in obsidian, however, are exceedingly rare, and in the obsidian flow under consideration it is confined to a small area, several hundred feet in extent, in that portion which poured into the old channel and acquired a greater thickness than that of the main flow.

A view of the west face of the southern end of the cliff is given in Pl. IX. The shining black columns rise from a talus slope which reaches some fifty feet up the cliff. These prisms are fifty or sixty feet high and vary in width from two to four feet near the end of the cliff, the width of each column being quite constant throughout its length. On the south face of this end of the cliff the columns are the same, but grow less clearly defined toward the east, where a sharp bend in the lava sheet has formed gaps in the rock and destroyed the continuity of the mass; beyond this the columns incline considerably toward the west, as though the underlying surface of contact sloped toward the west also. Farther up the slope to the east they disappear. The columns in the main face of the cliff are tilted 10° to the eastward, and the planes of flow which cross them have an average dip of 10° east, indicating that the underlying surface at this place slopes toward the east. Along the cliff to the north the columns become gradually broader, the largest being 20 feet in width.


The prisms have no uniform number of sides, four, five, and six being those most frequently observed; the sides are unequally developed, but at a distance the general effect is quite regular. Toward the north, with the change in the nature of the rock, to be described later on, the broad columns grade into massive blocks formed by vertical cracks much farther apart.

The rock forming the lower part of the columns is dense, black obsidian, with thin, lithoidal bands or layers of small spherulites, which are round, stony bodies, resembling hazel-nuts, or sometimes clay concretions. They have a radially fibrous internal structure and will be fully described (see p. 14). In this part of the columns there are almost no cavities or lithophysæ, as the hollow forms of spherulites are called, and but little contortion of the layers. Higher up, the rock is less massive and contains large lithophysæ flattened in the plane of flow. The tops of the columns pass into massive obsidian, which for 10 feet is quite dense, but above is full of large cavities which fairly honeycomb the mass. This may be seen on Pl. X, engraved from a photograph taken near the base of the columns, which, though nearly vertical, appear considerably inclined and distorted in the picture.


This upper portion, about fifty feet thick, is divided by vertical cracks into broad, quadrangular blocks having no resemblance to columns. The sides of the columns are comparatively straight and are independent of the flow structure within the mass, which is indicated by the microscopic banding of the obsidian and by the layers of spherulites which traverse the rock in parallel planes. The bending and twisting of these show the contortion of the viscous lava just before it came to rest. These layers pass through the columns at all angles, often showing abrupt folds and curves, which have been cut across sharply by the prismatic cracks. The crystalline layers formed planes of weakness through the rock, which has parted along them, producing transverse cracks that bear no fixed relation to the direction of the prisms, but follow the deviations of the flow. The contortion of the lava is particularly noticeable along the south face of the cliff, where the nearly horizontal layers in the most westerly columns curve upward and pass nearly vertically out of the present top of the cliff. Still farther to the east the layers are greatly twisted. At one place vertical rents and gaps between the layers show that the molten glass was so viscous and stiff before it finally came to rest that it pulled apart where the layers were vertical and did not close up again before the lava solidified.

The columnar portion on the west face extends for only a few hundred feet, the nature also of the rock changing in this direction. The lithoidal and spherulitic layers become more frequent until the black glass appears only in thin bands between light-gray layers, finally being represented by dark-gray lines between those of lighter color. This transition takes place in ascending the face of the cliff toward the northern end of the columnar portion, and also horizontally, the great bulk of the lava farther north, where the cliff is 200 feet high, being a light-gray, lithoidal rock, thinly fissile parallel to the planes of flow and bearing no resemblance to obsidian. In places through it black and red glass occurs in considerable quantity, mostly near the upper part of the cliff. The lithoidal form of the rock is not found in other parts of the sheet where it is one hundred feet and less in thickness, but only where it reached a much greater depth by filling up the old valley already mentioned. Unfortunately there is no indication of the original thickness of the lava at this place, as the top has been planed down by ice action and very considerably lowered. It is divided by two systems of vertical cracks into great quadrangular blocks which form a bold cliff rising 100 feet above the débris slope, a view of which is shown in Pl. XI.


As previously noticed, the columnar obsidian is only found in a small portion of that part of the lava that poured into a depression and acquired a greater depth than that of the flow in general, which is often 100 feet thick. The columnar portion was something over one hundred and sixty feet deep, but probably less than that of the lithoidal part farther north, where the mass cooled slowly enough to permit crystallization. The exceedingly rare occurrence of columnar structure in obsidian is probably owing to the fact that the conditions favorable for the production of prismatic structure and also for the solidification of the lava as amorphous glass are seldom coincident, the cause of columnar structure being unquestionably the shrinkage of a homogeneous rock which is cooling at a moderate rate from its surface. The very rapid surface cooling of a molten rock tends to produce cracks parallel to the surface,1 and cracks normal to the surface appear only in that portion of the mass which cools more slowly, the rate of cooling and the nature of the rock affecting the size of the columns, which are larger as the cooling is slower. The internal micro-structure of this obsidian and the fact that the spherulites and lithophysæ are cut sharply across by the planes of the cracks prove that the glass was rigid before the columns were developed. In general the obsidian shows but little tendency to crack, even where it has cooled rapidly, and in many parts of the flow no particular system of cracks is observed.

1The columnar structure in the igneous rock on Orange Mountain, New Jersey, J. P. Iddings: Am. Jour. Sci., 3d series, vol. 31, 1886, pp. 321-381.


If the cooling rock is not homogeneous in composition or texture the results of contraction will be altogether different. This is illustrated at the northern end of the cliff, where the lithoidal rock is traversed by widely distant, vertical cracks and a multitude of nearly horizontal ones which follow the planes of flow through all their complexities. The latter have resulted from differences in texture of the alternating layers, the most probable cause of which differences will be suggested later.

The origin of lamination or layer-structure frequently observed in many lava flows and so strikingly developed in the lithoidal portion of Obsidian Cliff is readily understood from the following: In a fluid free to flow over a horizontal surface the movement of the molecules will meet with least resistance in directions parallel to the plane of that surface; the fluid will therefore spread horizontally in all directions, producing a movement of its molecules in planes parallel to the underlying surface. Particles suspended in the fluid will be carried along these planes and portions differing in the amount or character of the suspended matter will be drawn out into layers along these planes of flow. In the case of lavas, the production of such layers will depend on the viscosity and lack of homogeneity of the mass at the moment of eruption and on the distance it flows. In the more liquid and homogeneous lavas, such as basalt, evidence of internal flow or lamination is less marked than in the more viscous and less homogeneous acid lavas, as rhyolite, where slight variations in the consistency of the mass find expression in bands and streaks of color or in layers of differing micro-structure and degree of crystallization. The greater the distance over which a viscous lava has spread, the thinner will be the layers of different consistency, which, near the source, may have been lenticular or quite irregularly shaped portions.

<<< Previous <<< Contents >>> Next >>>

Last Updated: 22-Jun-2009