Geological Survey Professional Paper 729—B Volcanic Stratigraphy of the Quaternary Rhyolite Plateau in Yellowstone National Park

Most of the Quaternary volcanic rocks exposed in Yellowstone National Park are related to the third eruptive cycle of the rhyolite plateau. Eruption and emplacement of these rocks were major geologic events of the middle and late Pleistocene history of the Rocky Mountain region. The key event in this cycle was eruption of the Lava Creek Tuff 600,000 years ago and collapse of the roof of its source magma chamber to form the gigantic compound Yellowstone caldera. The stratigraphic succession and petrologic evolution of this volcanic cycle must ultimately serve as a model for interpreting the more fragmentary record of the first and second cycles.

MOUNT JACKSON RHYOLITE

The Mount Jackson Rhyolite is defined here to include rhyolitic lava flows essentially conformable beneath the Lava Creek Tuff of the Yellowstone Group in areas near the wall of the Yellowstone caldera. The formation takes its name from the summit on the cliffs forming the north wall of the Madison Canyon 5-1/2 km west of Madison Junction (fig. 2). The Madison Canyon near Mount Jackson is the type area of the formation. The base of the Mount Jackson Rhyolite is not exposed in the type area, but the formation is overlain by the Lava Creek Tuff along the steeply sloping original surfaces of the flows. Little erosion occurred along this contact before emplacement of the Lava Creek, for the chilled glassy tops of the flows are everywhere preserved at the contact. Two rhyolite flows are present in the Madison Canyon area; the section there is as much as 450 m thick. The upper flow forms the summits and cliffs of Mounts Haynes and Jackson and is separated from the underlying flow by a thin fused tuff exposed near Harlequin Lake.

Rhyolite flows situated similarly with respect to the Lava Creek Tuff are exposed in the vicinity of Big Bear Lake, south of the Madison Canyon; near Wapiti Lake, west of the Mirror Plateau; on Flat Mountain, southwest of Yellowstone Lake; and in one small outcrop on Stonetop Mountain, north of Yellowstone Lake. Each of these flows is correlated with the Mount Jackson Rhyolite, and Boyd (1961, p. 392-393) included most of these localities in his Jackson flows.

The Mount Jackson Rhyolite flows are quite varied in general appearance, as is typical of rhyolitic lavas, but all contain 30-50 percent large phenocrysts of quartz, sanidine, plagioclase, and minor opaque oxides and clinopyroxene. These flows lithologically resemble younger rhyolitic lava flows of the third volcanic cycle and are distinguished mainly by their stratigraphic and structural positions.

J. D. Obradovich (written commun., 1970) has dated sanidine from the two flows in the Madison Canyon area as about 790,000 and 640,000 years old. This accords with the magnetic polarity time scale (Cox, 1969, fig. 4); remanent polarities of the two flows are, respectively, reverse and normal.

UNDINE FALLS BASALT

Basalts underlie the Lava Creek Tuff conformably in two areas. These basalts are named here the Undine Falls Basalt for the waterfall on Lava Creek near the Mammoth-Tower Junction Road just south of Mount Everts. The cliff on the north side of Lava Creek canyon just west of the falls is designated the type locality (fig. 2). The section there is about 6 m thick and consists of several thin basaltic flows overlying Cretaceous shales and sandstones and underlying the Lava Creek Tuff. Similar basalts, regarded as the same sequence of flows, occur as far northeast as Oxbow Creek and as far southwest as Obsidian Creek south of Winter Creek.

Probably correlative basalts occur north of Broad Creek, east of the Grand Canyon of the Yellowstone. Basalt, judged from mapping to be a single flow, occurs in that area between a flow of the Mount Jackson Rhyolite and the Lava Creek Tuff. The source vent is marked by a small scoria and cinder cone 3 km north of the Broad Creek crossing of the Wapiti Lake Trail. We believe this basalt north of Broad Creek to be at least approximately correlative with the type Undine Falls, inasmuch as both are conformable beneath the Lava Creek Tuff and have normal remanent magnetic polarities. We thus presume them both to be younger than 700,000 years.

PLATEAU RHYOLITE

Boyd (1961, p. 403-409) called the rhyolitic lavas younger than his Yellowstone Tuff, other than a few small rhyolite domes and mixed-lava complexes, the Plateau flows. He did not define them formally, but he clearly regarded them as a formation and named them as a composite of the flows of the Madison, Pitchstone, Central, and Solfatara Plateaus and nearby areas. We here name and define the Plateau Rhyolite in an essentially parallel way to include the rhyolitic lava flows of the Yellowstone rhyolite plateau that are younger than the Lava Creek Tuff. Inasmuch as we designate six new members of the formation, each with its own type area, the Plateau Rhyolite is defined here as the rock units that constitute the six named members.

Mallard Lake Member

The Mallard Lake Member of the Plateau Rhyolite is defined here to include rhyolite erupted within the Yellowstone caldera after it formed but before the rise within it of its two resurgent structural domes. The member consists of a single rhyolitic lava flow at least 120 m thick. It is named for a small lake near the center of its outcrop area, about 5 km northeast of Old Faithful (fig. 2). The area around Mallard Lake is designated the type area. The base of the member is nowhere seen in outcrop, but a drill hole near Rabbit Creek, just beyond the edge of the flow at the northwest end of the Mallard Lake resurgent dome, penetrated the underlying Lava Creek Tuff at a depth of only 10 m (White and others, 1969). The Mallard Lake flow is broken by northwest-trending normal faults that form a complex graben along the axis of the dome. The Mallard Lake Member and the graben system are overlain by the Upper Basin and Central Plateau Members of the Plateau Rhyolite. No complete sections of the Mallard Lake Member are exposed, and most of the original glassy shell of the flow has been eroded away. Most exposures are crystallized rhyolite with 20-30 percent phenocrysts of quartz and sanidine. Most mafic phenocrysts are altered to iron oxides. Flow layering is well displayed, and the rocks commonly are thinly flow-laminated.

The Mallard Lake Member is younger than the 600,000-year-old Lava Creek Tuff and is overlain by a flow about 530,000 years old (J. D. Obradovich, written commun., 1970). Structural reasoning suggests that the Mallard Lake is only very slightly younger than the Lava Creek.

Upper Basin Member

The Upper Basin Member, named here, is defined to include early postcaldera rhyolitic lava flows and associated pyroclastic deposits that were erupted within the Yellowstone caldera after resurgent doming. The Upper Basin Member is characterized by abundant plagioclase phenocrysts, commonly oligoclase that is highly embayed.

The type area of the Upper Basin Member is in the Upper Geyser Basin (fig. 2), where outcrops of the member are scattered widely among the sediments and hydrothermal deposits that floor the basin. Four drill holes in the Upper Geyser Basin all bottomed in rhyolite of the Upper Basin Member. These outcrops and drill cores from the Upper Geyser Basin, as well as many outcrops north along the Firehole River nearly to Tangled Creek in the Lower Geyser Basin, are interpreted to belong to a single flow. The steep front of a younger flow of the member forms the south end of the Upper Geyser Basin; the same flow forms the topographic surface around Scaup Lake. Both flows in the type area of the member overlie the Mallard Lake Member and underlie the Central Plateau Member of the Plateau Rhyolite.

Other than in the geyser basins of the Firehole River, rhyolites of the Upper Basin Member are exposed only in the upper Grand Canyon of the Yellowstone and in the area southeast from it to Fern Lake. Boyd (1961, p. 405-406) discussed the Canyon flow, named for the Grand Canyon of the Yellowstone, and relations of that flow along the rim to pyroclastic rocks deeper in the canyon. The flow and tuffs seem to be gradational from one to the other, which led Boyd to suggest tentatively that the Canyon flow is an example of a froth flow, an eruptive mechanism proposed by Kennedy (1955, p. 495) to explain some features of welded tuffs. We disagree with this explanation, but discussion of the mechanisms of eruption is deferred to a later paper. In our interpretation, the rhyolite section in the upper Grand Canyon area consists of a basal, largely agglutinated air-fall tuff overlain successively by the Canyon flow proper and by a younger flow exposed near the Dunraven Pass Road.

The older Mallard Lake Member and the younger intracaldera members of the Plateau Rhyolite generally have little or no phenocrystic plagioclase. The Upper Basin Member, by contrast, has considerably more plagioclase than sanidine phenocrysts, and in some flows the feldspar phenocrysts are exclusively plagioclase. Clinopyroxene too is considerably more abundant in the Upper Basin than in the other members.

Potassium-argon dating, by J. D. Obradovich (written commun., 1970), of sanidine, plagioclase, and nonhydrated glass from rhyolites of the Upper Basin Member shows that most of the flows are only slightly younger than the Lava Creek Tuff, ranging from about 600,000 to 530,000 years old. One flow, however, which we correlate on the basis of its phenocryst mineralogy with the Upper Basin Member, is about 260,000 years old.

Obsidian Creek Member

The Obsidian Creek and Roaring Mountain Members of the Plateau Rhyolite each comprise a series of isolated rhyolitic lava flows and domes that lies wholly outside the Yellowstone caldera and is, therefore, particularly difficult to define in traditional stratigraphic terms. In a sense they are extracaldera parallels of the intracaldera Upper Basin and Central Plateau Members. With this parallel in mind, we here define the Obsidian Creek Member as early post-Lava Creek rhyolitic domes and flows outside the Yellowstone caldera in which plagioclase is a common phenocrystic constituent. In terms both of age and of the petrography of their rhyolitic components, two lava flows of mixed rhyolite and basalt can reasonably be included in the member. The following bodies constitute the member (and thus, by definition, its type area): The mixed lavas of Gardner River (Fenner, 1938, 1944; Wilcox, 1944; Hawkes, 1945) and Grizzly Lake (Boyd, 1961, p. 403); and the Willow Park, Apollinaris Spring, Landmark, Gibbon Hill, Geyser Creek, and Paintpot Hill eruptive domes (fig. 2). The mixed lavas and the domes occur in a north-trending belt between the Norris Geyser Basin area and the Gardner River. Obsidian Creek, for which they are named, parallels much of this belt.

The mixed lava of Gardner River lies partly between flows of the Swan Lake Flat Basalt and partly on the Lava Creek Tuff. All the other bodies can be observed, or can be inferred with considerable confidence from map relations, to lie on the Lava Creek. The mixed lava of Grizzly Lake forms the top of its local stratigraphic section, but the Gardner River, Willow Park, and Apollinaris Spring bodies are overlapped by flows of the Swan Lake Flat Basalt, and the Landmark, Gibbon Hill, Geyser Creek, and Paintpot Hill bodies are overlain by younger flows of the Plateau Rhyolite.

No radiometric age determinations have been made on the Obsidian Creek Member. Stratigraphic relations just cited, however, suggest that at least most of the member is only slightly younger than the Lava Creek Tuff.

Central Plateau Member

The Central Plateau Member of the Plateau Rhyolite is defined here to include late postcaldera rhyolitic lava flows, some of them very large, erupted from vents within the Yellowstone caldera. Eighteen separate flows and domes have been mapped as parts of the member, but seldom can one find more than three flows superposed in a single stratigraphic succession. The type area is designated as the Central Plateau (fig. 2), located in the middle of the Yellowstone caldera between its two separately identified collapse segments. The flows of the member form the Madison, Pitchstone, Central, and Solfatara Plateaus and a few lower areas nearby. There may be other flows in the sequence that are entirely buried. The member constitutes by far the greatest part of the total volume of the Plateau Rhyolite.

The Central Plateau Member overlies the Obsidian Creek, Upper Basin, and Mallard Lake Members of the Plateau Rhyolite; the Lava Creek Tuff; and a few older units. Only surficial sediments of late Pleistocene age overlie the Central Plateau Member. Glacial and nonglacial sedimentary deposits have been identified locally within the member, as is noted in the final section of this report.

Some individual flows of the member are 300 m or more in thickness, and a few extend more than 20 km from their source vents. The flows have abundant phenocrysts, generally 30-50 percent. Most of them have mainly quartz and sanidine with no plagioclase; a few have minor plagioclase. All of them contain in their glassy portions minor clinopyroxene, opaque oxides, and fayalitic olivine.

Potassium-argon age determinations on sanidine from the Central Plateau Member range from about 200,000 to about 70,000 years (J. D. Obradovich, written commun., 1970).

Shoshone Lake Tuff Member

A single rhyolitic ash-flow tuff crops out within the Plateau Rhyolite sequence. We name it here the Shoshone Lake Tuff Member for exposures on the steep slopes west of Shoshone Lake. The type section is in the gully north of Fall Creek (fig. 2), west of the Shoshone Lake Geyser Basin, where the member is about 180 m thick and lies between two flows of the Central Plateau Member.

The member is widespread between West Thumb and Shoshone Lake. Elsewhere, the tuff is largely covered by younger rhyolite flows, but small exposures occur between various flows of the Central Plateau Member in the Bechler River area, near Little Firehole Meadows, and in the upper drainage of Nez Perce Creek. In the area of Shoshone Lake the ash-flow tuff has an irregular distribution which is hard to explain by erosion but which suggests the possibility that it was emplaced on a lobe of glacial ice that occupied the lake basin. The distribution and the pattern of welding and crystallization zones of the member indicate that its source area was in the vicinity of West Thumb, which may, in fact, be a caldera lake formed by collapse after eruption of the tuff.

Most of the type section consists of glassy nonwelded to partially welded tuff, but local vapor-phase and devitrified zones a few meters thick occur in the upper part. The degree of welding varies vertically between more and less densely welded tuff. By contrast, the Shoshone Lake Tuff Member near West Thumb, the deep western basin of Yellowstone Lake, commonly is more densely welded and contains a much thicker vapor-phase zone in proportion to glassy welded tuff. The basal part of the member is well exposed in the roadcut above Bluff Point, along the northwest shore of West Thumb. The basal glassy zone there is less than 5 m thick and rests against a flow of the Central Plateau Member. A specimen from the vapor-phase zone at the Bluff Point locality was illustrated by Ross and Smith (1961, fig. 11, p. 27). The upper part of the section at Bluff Point probably is repeated by faulting, but exposures are poor and do not allow determination of the stratigraphic thickness. The tuff appears to be overlain by a higher flow of the Central Plateau Member. About 2-1/2 km southwest of Bluff Point, in a gravel pit northwest of Potts Hot Springs, the Central Plateau Member overlies gravels containing boulders and cobbles of the Shoshone Lake Tuff Member.

The Shoshone Lake Tuff Member contains 20-30 percent phenocrysts of quartz, sanidine, subordinate plagioclase, and minor clinopyroxene and opaque oxides. It contains more abundant pumice than tuffs of the Yellowstone Group and generally has a coarser matrix. Inclusions of older rhyolite, mainly black vitrophyre, are abundant throughout the member.

The age of the Shoshone Lake Tuff Member is bracketed by the ages of two flows of the Central Plateau Member, sanidine from which has been dated as about 200,000 and 150,000 years old respectively; sanidine from a bedded tuff in the Potts Hot Springs gravel pit, probably a part of the Shoshone Lake, is 180,000 years old (J. D. Obradovich, written commun., 1970).

Roaring Mountain Member

We name the Roaring Mountain Member here for four young phenocryst-free or phenocryst-poor extracaldera rhyolitic lava flows that occur as isolated bodies north of the Yellowstone caldera. The member comprises the Crystal Spring, Obsidian Cliff, Cougar Creek, and Riverside flows (which, by definition, constitute its type area, fig. 2); Roaring Mountain is in the vicinity of the Crystal Spring and Obsidian Cliff flows, which occur in the same belt as the Obsidian Creek Member. The Cougar Creek and Riverside flows are in the area west and northwest of Madison Canyon. The four flows overlie the Lava Creek Tuff, the Swan Lake Flat Basalt, and parts of the Obsidian Creek Member of the Plateau Rhyolite. The Cougar Creek and Riverside flows are overlain by flows and cut by dikes of the Madison River Basalt. The other flows are overlain only by surficial deposits of late Pleistocene and Holocene age.

All the flows are phenocryst-free or very phenocryst-poor rhyolites which contain abundant fresh black obsidian as well as crystallized and partly crystallized material. Phenocrysts, where present, are quartz, sanidine, plagioclase, and opaque oxides.

Obsidian from three of the four flows has been dated by the K-Ar method by J. D. Obradovich (written commun., 1970); the dates seem to indicate a considerable age span for the member. The Crystal Spring and Obsidian Cliff flows yielded young apparent ages (about 160,000 and 75,000 years respectively); the Cougar Creek flow gave an age of 400,000 years. However, the reliability of these obsidian dates is uncertain.

POST-LAVA CREEK BASALTS

Basalts younger than the Lava Creek Tuff occur around the margins of the rhyolite plateau surrounding the caldera. Because different relations are displayed in several widely separated areas, these basalts have been grouped into five named formations.

Swan Lake Flat Basalt

Basalts younger than the Lava Creek Tuff but older than an unconformity that represents the first major period of canyon cutting into the Lava Creek occur widely in the area north of Obsidian Cliff, nearly to Mount Everts. We here name these lavas the Swan Lake Flat Basalt after a valley partly surrounded and largely underlain by them about 8 km southwest of Mammoth. The type area (fig. 2) is designated as the Sheepeater Cliffs, on the east side of the canyon of the Gardner River southeast of Bunsen Peak. Flows of the lower part of the sequence are well exposed there and are conformable on the Lava Creek Tuff, whose easily eroded nonwelded glassy cap is locally preserved. Flows of the Plateau Rhyolite locally overlie the Swan Lake Flat Basalt. No single section, however, exposes well both the base and the top of the formation. East of the Sheepeater Cliffs, basalts form a rounded hill that probably marks one of the source vents for the flows; the formation totals nearly 200 m in thickness in that area. Other cinder cones that probably mark vents for the Swan Lake Flat Basalt occur near Horseshoe Hill, northeast of Obsidian Cliff.

Boyd (1961, p. 402) observed that basalts of two ages occur in the Bunsen Peak area, one conformable on the Yellowstone Tuff (our Lava Creek Tuff) and the other filling a younger canyon. We recognize the former as the Swan Lake Flat Basalt and the latter as the Osprey Basalt.

The Swan Lake Flat Basalt generally is light gray to moderate gray and contains sparse to abundant large phenocrysts of plagioclase, as much as 1 cm across, and locally some olivine phenocrysts. Outcrops of a lithologically distinctive basalt in the area near Geode Creek (Howard, 1937, p. 19-21) probably represent a mixed lava dominated by basalt and have therefore been mapped separately. A faint streakiness, a few xenocrysts, and considerable textural heterogeneity show the mixing in hand specimen. Extremely large plagioclase phenocrysts, as much as 5 cm across, occur rarely in the rock. At least one flow near Geode Creek appears to have been fed through a vent now preserved as a low scoria mound above an exposed dike about midway between the main Mammoth-Tower Junction Road and the unpaved Crescent Hill Road, 1-1/4 km east of Geode Creek.

The age of the Swan Lake Flat Basalt is suggested by its stratigraphic relations. The Swan Lake Flat is conformable on the middle Pleistocene Lava Creek Tuff, which probably is somewhat younger than the Cedar Ridge (Kansan) Glaciation of the Rocky Mountains (Richmond, 1970b, p. 8, 21). The Swan Lake Flat is older than canyon cutting that predated the Osprey Basalt. Post-Lava Creek canyon cutting in the drainage of the Yellowstone River predated a till of probable Sacagawea Ridge (Illinoian) age in the Grand Canyon area (Richmond, 1970a; 1970b, p. 21). Thus the episode of canyon cutting between the Swan Lake Flat and Osprey Basalts probably was contemporaneous with the Yarmouth Interglaciation of the Great Plains. Richmond (1957) noted that the episode of deepest canyon cutting during the Pleistocene occurred in many parts of the Rocky Mountains during this same interglaciation.

Falls River Basalt

The Falls River Basalt is named here to include basalts younger than the Lava Creek Tuff and older than the Central Plateau Member of the Plateau Rhyolite in the region of the southwest corner of Yellowstone National Park. The formation takes its name from the major stream draining that area. The type locality is the west rim of the valley of Falls River just below Cave Falls (fig. 2), where the formation overlies the Lava Creek and is overlain by upper Pleistocene glacial deposits. In the drainage area of the Bechler River about 8 or 9 km north of its confluence with Falls River, just above Cave Falls, the Falls River Basalt is overlain by several flows of the Central Plateau Member.

The Falls River Basalt is very similar in lithology to the Swan Lake Flat Basalt, commonly containing sparse phenocrysts of plagioclase about 1/2-1 cm across. The basalts generally are moderate gray and form dense thin flows with vesicular tops. Just west of the southwest corner of the national park these flows make up a low shield, Rising Butte (fig. 2), which probably was the source of at least some of the flows in the park.

The age of the Falls River Basalt is bracketed by the ages of the 600,000-year-old Lava Creek Tuff and the overlying Plateau Rhyolite flows, probably about 100,000 years old. Apparent conformity of the Falls River and the Lava Creek suggests that the basalt is closer in age to the underlying tuff than to the overlying lavas.

Basalt of Mariposa Lake

A few scattered outcrops of basalt southeast of the rhyolite plateau are discussed in this section even though their stratigraphic position relative to the plateau sequence is unknown. These outcrops of basalt all occur on Two Ocean Plateau and are underlain by Eocene rocks of the Absaroka Volcanic Supergroup. These basalts have been mapped by our colleagues H. W. Smedes and H. J. Prostka, and the following discussion is based mainly on their observations. The basalts are as thick as 50 m and lie on an erosional surface of considerable relief. They are overlain by upper Pleistocene glacial deposits. Their age is unknown, but their lithologic and general petrographic characteristics are quite unlike those of the Absaroka volcanic rocks and much like those of upper Cenozoic basalts of the region. They are very similar to the Swan Lake Flat and Falls River Basalts in lithology and in their spatial relations to the rhyolite plateau except that no ash flows of the Yellowstone Group, our regional stratigraphic framework, were emplaced on the surface of Two Ocean Plateau. These relations suggest the possibility that the basalts of this area are stratigraphic relatives of other post-Lava Creek basalts, but the basalts on Two Ocean Plateau could be older.

The upper Cenozoic basalts of Two Ocean Plateau are designated here as the basalt of Mariposa Lake; they partly fill an old valley in which that lake lies (fig. 2). These basalts also occur on the flat ridge about 4 km northwest of Mariposa Lake and in two patches on the ridge west of the Yellowstone River between Badger and Phlox Creeks. Other isolated patches may be present beneath the widespread mantle of glacial debris in the area.

Madison River Basalt

Post-Lava Creek basalts which occur in widely scattered patches in the area south and west of the Gallatin Range to the vicinity of the Madison Canyon and the eastern part of the Madison Valley near West Yellowstone are named here the Madison River Basalt. We designate the basalt-covered up lands west of the south end of the Gallatin Range, south of upper Maple Creek, the type area. The lavas lie on an eroded surface on the Lava Creek Tuff but predate deeper canyon cutting in the area. Locally they overlie the Cougar Creek and Riverside flows of the Roaring Mountain Member of the Plateau Rhyolite. Some outcrops in the type area are deeply mantled by upper Pleistocene glacial deposits, and till of Bull Lake (early Wisconsin) age south of West Yellowstone contains abundant clasts of Madison River Basalt. Some flows are so small that it is difficult to show them on maps at a scale of 1:62,500. Other outcrop areas contain two or more flows, cover several square kilometers, and represent sections more than 50 m thick. Like the Swan Lake Flat and Falls River Basalts, the Madison River Basalt contains moderately abundant plagioclase phenocrysts and relatively rare olivine phenocrysts, and it generally is moderate gray to light gray. Chemically, however, these basalts are different. Although the chemistry of the rhyolite plateau volcanic rocks has not been used in defining the formations, it is worthy of note that the other basalts of the region are all olivine tholeiites with very low potassium contents. The Madison River Basalt, however, contains markedly more potassium, iron, titanium, phosphorus, fluorine, rubidium, lead, and uranium and contains less calcium and magnesium, even when compared with the other basalts of the area of similar silica content.

The Madison River Basalt is at least partly younger than the Roaring Mountain Member of the Plateau Rhyolite. Relations just noted suggest that the Madison River is younger than at least part of the Swan Lake Flat and Falls River Basalts, but it is entirely older than glacial deposits of early Wisconsin age and may be largely older than the canyon cutting contemporaneous with the Yarmouth Interglaciation. (See discussion of the age of the Swan Lake Flat Basalt.)

Osprey Basalt

The Osprey Formation was defined by Pierce, Christiansen, and Richmond (1970) to include basalts and minor interlayered gravels in the northern Yellowstone National Park area that are younger than the Swan Lake Flat Basalt. Subsequent work by K. L. Pierce and by us has shown inclusion of the Tower Creek Gravel Member in the Osprey to be erroneous. As a result, we now exclude the sediments and basalts of The Narrows and change the name from Osprey Formation to Osprey Basalt.

The flows and interlayered gravels of the Osprey Basalt partly fill deep canyons eroded into the Lava Creek Tuff and the Swan Lake Flat Basalt. As discussed in the section on the Swan Lake Flat Basalt, the valleys in which the Osprey was deposited may be of the same age as the Yarmouth Interglaciation; the formation is overlain by glacial deposits of late Pleistocene age.