University of Washington Publications in Geology The Geomorphology and Volcanic Sequence of Steens Mountain in Southeastern Oregon

STEENS MOUNTAIN BASALT
(continued)

AGE RELATIONSHIP

From fossil leaves, the Upper Cedarville Beds beneath the Warner Basalt were considered by Chaney¹⁸ to correlate with the Mascall Formation overlying the Columbia River Basalt in the John Day Valley. They were, therefore, thought to be of Middle Miocene age. Vertebrate remains, however, found in the same beds indicated a later age. From paleobotanical evidence the Upper Cedarville beds should correlate with those of Alvord Creek at the base of Steens Mountain, but stratigraphically the entire Steens Mountain volcanic series apparently lies between them. Although the extrusion of these varied volcanic rocks probably was relatively rapid, the evidence indicates that the Mascall flora was quite persistent and possibly that it continued into the Pliocene.

---

¹⁸R. J. Russell, op. cit., p. 412.

---

To the east Russell definitely correlates¹⁹ the Warner Basalt with the Mesa Basalt and tentatively not only with the flow exposed to the north at Railroad Ridge, but also "at least in part" with the Pueblo Mountain series. The Mesa Basalt caps the Virgin Valley Beds which contain in their lower horizons vertebrates of the Middle Miocene.²⁰ These lower beds should therefore be older than the Steens Mountain volcanic series. The upper strata, however, appear to be nonfossiliferous, so that the precise age of the Mesa Basalt cannot be determined. The Railroad Ridge Basalt caps the Thousand Creek Beds, which Stock²¹ considers from a study of their vertebrate fauna to be fairly late in the Pliocene. These beds, however, in the opinion of the writer overlie the Pueblo Mountain series approximately conformably. This series, as previously mentioned, appears from almost continuous exposures to be the direct stratigraphic continuation of the Steens Mountain lavas.

---

¹⁹Op. cit., pp. 416-417.

²⁰Chester Stock, personal communication.

²¹Ibid.

---

Since the basalt forming these three units at the southern end of Pueblo Mountain is not sufficiently distinctive to permit a petrographic correlation, the problem at present cannot be definitely settled, but it appears certain that the Pueblo Mountain series and the Railroad Ridge Basalt are widely separated stratigraphically. Locally, however, no definite evidence was obtained to prove whether the Mesa Basalt correlated either with the Railroad Ridge lava, which it resembles physiographically, or with the uppermost lava of the Pueblo Mountain series. The basalt exposed near the southern end of Pueblo Mountain, as well as being overlain by acidic lava and tuffaceous beds, has been intruded by a number of small sills of obsidian. Although, possibly due to this later activity, the series is quite badly altered, the effect may be too local to be used as a distinctive criterion. The facts suggest, however, that these flows have an earlier origin than the Mesa Basalt. The latter, therefore, may correlate with the Warner Basalt as Russell proposed.

In general the evidence indicates that the Steens Mountain Basalt is definitely younger than the Columbia River flows and occurred either late in the Miocene or early in the Pliocene. In the Warner Valley region the overlying tuffaceous beds, which are presumably a continuation of the Upper Cedarville, suggest that the Mascall flora continued into the Pliocene, as the vertebrate evidence indicates. The Warner Basalt, overlying these beds slightly disconformably, appears to thicken southward rather than northward as Russell suggested. On the other hand the apparent absence of the Steens Mountain series both in Virgin Valley and on Warner Range indicates that it ceases abruptly to the south, but the actual relationship to its barrier was not determined. In Virgin Valley, however, the advance of these lavas may have been halted by the great thickness of the Canyon Rhyolite, which lies on the northern side.

PETROGRAPHY

Although most of the basalt of southeastern Oregon is normal in its mineral content, it exhibits a number of textural and structural characteristics, which easily distinguish it from the more typical varieties encountered to the north on the Columbia River Plateau. Some of the most distinctive regional characteristics, however, are repeated at such widely separated horizons that they are not of any great stratigraphic value. Petrographic correlation is also complicated by the fact that such striking features as the concentration of phenocrysts may vary rapidly in a single flow.

The extreme fluidity of the lava has already been indicated by the remarkable thinness of the flows, their level surfaces, and the merging of successive sheets. This testimony is also supported by petrographic evidence, for most of the textural peculiarities demand great mobility, which was presumably caused by an unusually high volatile content.

Although the series is composed predominantly of fairly basic olivine basalt, the perfectly fresh rock is of a far lighter shade of gray than one would expect. Most of the flows, in spite of their thinness, are relatively coarsely holocrystalline. As a rule the texture is almost uniform throughout each sheet. In many flows, the most noticeable megascopic constituents consist of coarse plates of clear glassy labradorite and small grains of olivine, which, as a rule, has been rendered iridescent by a thin film of alteration. In most flows, however, the texture of the groundmass forms the most unusual feature.

Under the hand lens, it is clearly apparent that open cavities exist locally between the crystalline constituents. The most striking characteristic of these minute irregular cavities is the presence of delicate laths of light grey labradorite, which project into them, and in some specimens, form a network of conflicting plates. These usually range from .5 to 1.5 mm. in length. Less commonly olivine and augite come in contact with these spaces. The angular shape of the cavities resembles the interstitial areas occupied by the glassy mesostasis in some basalts.

The only established term that suggests this texture is "miarolitic," but this has been previously applied to plutonic rocks and as a rule in reference to the presence of drusy cavities lined with a crystalline coating. In this basalt, however, the cavities are not of the drusy type, and the orientation of the enclosing crystals has no reference to the space they bound. Since this texture apparently has not been previously described, the writer proposes the name "diktytaxitic" (Greek, diktuon, net, + taxis, arrangement)²² in reference to the net-like arrangement of the feldspar laths.

---

²²The writer gratefully acknowledges the advice of Dr. Charles E. Weaver and Professor G. F. Goodspeed in the selection of this name.

---

"Diktytaxitic" texture is most clearly defined in coarsely crystalline rocks, where the cavities are usually relatively uniform in their distribution. In this type, however, the examination is necessarily confined to a hand lens (fig. 71), for most of the specimens are too friable to preserve their features in thin section, except when the cavities have been subsequently filled with zeolites such as chabazite or thompsonite, which have been derived from the endomorphic alteration of the feldspar. In the more aphanitic varieties, which are usually darker in color, the open spaces are more scattered in their distribution. The consolidation of this type is usually sufficient to permit the grinding of satisfactory thin sections (fig. 72).

Fig. 71. Photograph (magnified 7X) of a hand specimen of a basaltic flow, which was observed near the base of the series in the valley of Mann Creek. The flow, which has a maximum thickness of about 70 feet, exhibits the coarsest crystallization observed. It shows typical "diktytaxitic" texture with labradorite laths forming an open network of plates. The olivine occurs as clear light colored grains, while the ophitic augite is dark. The deeper cavities appear black.

Fig. 72. Microphotograph of a thin section of a basaltic flow exposed at the northern end of Hart Mountain at the top of the lower series. "Diktytaxitic" texture is rendered more apparent by the use of the mica plate and crossed nicols. The cavities are white. The visible mafic consists principally of partially altered grains of olivine. (36X).

The distribution of the labradorite and the size of the laths appear as a rule to he relatively uniform throughout the groundmass of an individual specimen. In part, however, the plagioclase is intergrown with ophitic augite. This mineral occurs generally in isolated patches composed of one or more crystals, and is usually enclosed by a zone containing intersertal grains of olivine and magnetite. In fact, in some specimens the intersertal grains have been completely converted to iron oxide, which renders the ophimottling very apparent in thin section. These mafics are seldom enclosed in the augite.

Ophimottling is probably the most persistent feature in the texture of these basalts. It is also encountered in this region in many rocks that show no tendency of being "diktytaxitic." Some of the more aphanitic flows, however, show small intersertal grains of augite as the principal mafic, while olivine may be either absent or a very minor constituent. A few rocks show the feldspar laths imbedded locally in an almost opaque ground, which suggests decomposed glass. As a rule this substance shows parallel streaks principally defined by magnetite. This type is formed from the alteration of a mafic mineral that exhibits a single well defined cleavage, and an extinction that suggests augite.

As a rule, however, the augite is free from alteration while the olivine is found in various stages of decomposition. In some instances the latter appears to have been discolored to a reddish brown without being completely decomposed. The staining usually is marginal, although in a few specimens the centers of the crystals are discolored while the margins are perfectly fresh. More often the olivine is found to be altered to a dark orange-brown non-pleochroic platy mineral which exhibits parallel extinction with the same orientation as the olivine. This product is classed as iddingsite.

In some specimens, the alteration is confined to these clean-cut pseudomorphs of olivine. In others, the decomposition of this mineral has contributed to the formation of both green and orange colored deuteric residuals, which locally exhibit fibrous structure possibly due to the crystallization of chlorite. These substances, which are at least similar to chlorophaeite, locally fill the "diktytaxitic" cavities and even intrude the cracks in the feldspar. The rock thus impregnated is of a very dark color and forms a marked contrast with the unaltered facies. In a number of flows a "sheaf-and-core structure"²³ is defined by a localization of this type of alteration in the center of the major joint blocks. A discussion of the origin of this remarkable feature will be treated in detail in a later paper.

---

²³"Tertiary and Post-Tertiary Geology of Mull," Memoirs of Geol. Surv. Scotland, pp. 261-264, 1924.

---

The phenocrysts of olivine are usually less than 2 mm. in diameter, but in some flows their mass was sufficient to permit gravitational accumulation in the very fluid lava. The phenocrysts of plagioclase are far more variable in size. They occur principally as plates, which locally attain a diameter of over 4 cm. In many flows they are relatively uniform in their concentration and show no tendency to alignment. In this type a radial grouping of the coarse crystals is quite characteristic (fig. 73). In some flows, however, they are very variable in their concentration and occur either as swirls or in roughly horizontal zones. Some of these variations are attributed to the merging of successive sheets of slightly different composition, while others are considered to be due to gravitational action. These petrogenetic features as well as the consideration of the origin of the horizontal segregations, which are quite common in the Steens Mountain Basalt, are also left to a subsequent publication.

Fig. 73. A basaltic flow near the base of the series to the south of Alvord Creek, showing the typical radiating clusters of rather small plates of basic labradorite. A 3/4 inch disk furnishes the scale.

The plagioclase which in most flows forms about 60 per cent of the mineral content, shows little or no tendency towards zonal growth. The composition appears to range between an intermediate and a calcic labradorite with a preponderance of the latter. Most of the crystals of the groundmass, as well as the phenocrysts show a distinct crystal outline. In a number of specimens, however, feldspar also occurs locally as interstitial grains. These poorly defined grains invariably show an indistinct wavy extinction and an index of refraction which is but slightly less than the crystals which they enclose.

Many of the ill-defined sheets near the base of the series are amygdaloidal, with thompsonite and chabazite filling amoebaform vesicles. As their composition suggests, these zeolites are derived from the decomposition of the plagioclase. This alteration is considered to be endomorphic and to have been caused by the retention of the volatiles beneath the overlying sheets which had advanced prior to solidification. Locally, however, calcite and both chalcedonic and crystalline quartz are found either as amygdules or as a partial filling of irregular cavities between the flows. Scolecite was observed in association with the calcite. This deposition, at least in part, has probably been caused by later hydrothermal activity. In addition hyalite is quite commonly encountered as a coating on the major joint cracks.

Chemical analyses of this basalt (table VIII) indicate that it is fairly constant in composition, although at least some of the upper flows are slightly more acidic. Flows of a similar type, however, occur sporadically throughout the series. The specimens analyzed fall predominantly in the subrang of Hessose, which demands a moderately high content of lime. Although typical basalts, they show a far higher content of alumina and titanium oxide than usual. Since these constituents should not be expected to add to the fluidity of the flows, their physical characteristics are all the more remarkable. In the calculation of the norms, the high content of ilmenite and magnetite, and in some cases hematite, depletes the ferrous iron and thereby results in the development of residual quartz, in what are in reality fairly basic basalts.

TABLE VIII

The alumina is explained by the unusually high feldspar content, which causes the rock to approach an effusive equivalent of anorthosite. It is not determined, however, in what mineral the titanium occurs. The augite as a rule is colorless or slightly brownish in thin section and does not suggest a titaniferous variety. A partial analysis of some of the olivine which is not included, shows only a trace of titanium dioxide. It is therefore considered to be present as ilmenite, as the norm suggests, although the characteristic alteration to leucoxene was not observed.

BASALTIC INTRUSIONS

Basaltic dikes are frequently encountered on the eastern scarp of High Steens. A number of these may be observed to traverse the underlying tuffs and lavas, but some of the most prominent ones are exposed cutting the basaltic series in the lower cirques. Both the size and the orientation of these dikes vary greatly. Most of them, however, are vertical and the larger ones are confined to an approximate north and south axis (fig. 6).

Many of these dikes show large phenocrysts of labradorite and presumably form the feeders for the overlying basaltic series, which they closely resemble. Commonly, however, they lack the open texture, and resemble the altered phases of the flows, for the retention of volatiles has resulted in the impregnation of the rock with deuteric residuals. In some instances the alteration is very irregular and in other cases confined to well-defined vertical bands. In some dikes, chlorophaeite occurs as minute globules, which are usually concentrated in the vertical bands. The marked contrast in the resistance of the two facies may give the appearance of a multiple dike. The localization of the alteration appears to be due to the irregular distribution of the volatiles at the time of consolidation. The alignment may be explained by marginal chilling, but the evidence is not conclusive.

A number of the larger dikes are formed by successive intrusions and hence are true multiple dikes. The largest of these multiple dikes is well exposed near the base of several of the lower cirques and on the intervening divides. It may be traced for many miles as a wall which is usually about 50 feet in width and locally rises close to 100 feet in height (fig. 74). Although the exposures are not continuous, Waring²⁴ considered the prolongation of this large dike to form a prominent exposure near the southern end of the eastern scarp bounding Smith Flat.

---

²⁴G. A. Waring, "Geology and Water Resources of the Harney Basin Region, Oregon," U.S. Geol. Survey, Water-Supply Paper 231, p. 22, 1908.

---

Fig. 74. On the left may be seen the broad multiple basaltic dike cutting the basaltic series in the cirque of the south fork of Alvord Creek. The saddle marks the position of the 200 foot normal fault, which dips westward at about 80°. The prominence on the right rises about 500 feet above the saddle. The flows exposed to the west have been bevelled by sliding talus.

In spite of the numerous basaltic dikes exposed on the eastern scarp of High Steens an actual locus of extrusion was observed only at the base of the Mosquito Creek cirque on its northern side (fig. 75). Unfortunately, the locality, especially to the east, is covered either with talus or with morainal material, which renders it impossible to determine if the vent were formed by a true fissure eruption. The exposure shows a coarse basaltic breccia enclosed in a reddish vesicular lava, some of which is of a brilliant shade. At its western margin, this lava crosscuts the previously mentioned bed of grey stratified tuffs, which is interbedded near the base of the series. Upwards, to the west, the basalt merges into thin flows, some of which are less than a foot in thickness (fig. 76). Jointing and differential weathering causes these flat sheets to be very apparent, but on examination the actual contacts are locally difficult to determine. These sheets are considered to have been formed by successive gushes of lava that was so fluid that it was able to attain a relatively level surface and to become approximately stationary before the next sheet submerged the partially solidified surface.

Fig. 75. The flaring margin of the basaltic vent crosscutting the stratified tuffs interbedded with the lower basaltic flows on the northern side of the Mosquito Creek cirque.

Fig. 76. Thin basaltic sheets extruded from the vent in the Mosquito Creek cirque.

Although most of the basaltic dikes are approximately vertical, two small ones, in the fault zone between Toughey Creek and Pike Creek, follow an inclined course parallel to the planes of the step faults and dip eastward at about 60°. The exposures are not sufficiently distinct to prove if either of the dikes were displaced. It is probable that these dikes were either contemporaneous with or subsequent to the initial movement on the faults, and therefore far later than the main basaltic series. One of these dikes, on coming in contact with a tuffaceous bed, follows it for about 100 feet and then cuts upwards with the same inclined orientation as its lower part. These dikes are probably contemporaneous in origin with the small sills of similar composition, which intrude the lower tuffs in Pike Creek.

The large basic sill exposed in the valley of Indian Creek, beneath the southern continuation of the upper flow of platy rhyolite, is also considered to be of late origin. Its precise relationship has not been determined, but it is remarkably fresh and shows no resemblance to any of the lavas observed on the mountain. Megascopically its central facies suggests a coarsely holocrystalline rock. Petrographically it consists of about equal quantities of sodic labradorite and pale brownish augite in a rather opaque grayish or brownish ground, which forms about 30 per cent of this facies. This ground contains a high content of magnetite, chiefly in thin skeleton plates, which attain locally a diameter of nearly 4 mm. A few laths of plagioclase are also over 3 mm. in length, although they average approximately 1 mm., and are but slightly coarser than the augite with which they are associated. Most of the plagioclase and some of the pyroxene show euhedral faces. The minerals exhibit no predominant alignment and no tendency to ophitic intergrowths. The principal accessory appears to be an indeterminable zeolite, which has been partially altered.