Phlaocyon leucosteus, Matthew, 1899

HUNT, ROBERT M., 2005, An Early Miocene Dome-Skulled Chalicothere from the ‘‘ Arikaree’ ’ Conglomerates of Darton: Calibrating the Ages of High Plains Paleovalleys Against Rocky Mountain Tectonism, American Museum Novitates 3486, pp. 1-46 : 32-41

Phlaocyon leucosteus

Figure 23

A left maxilla with P3­M2 (UNSM 44822) belonging to a small canid is attributable to Phlaocyon , a hypocarnivorous borophagine lineage noted for the addition of accessory cusps to its P4 and molars. It was found in a sand­and­gravel lens at East Sturdivant Butte at the same stratigraphic level as Aletomeryx (UNSM 44820, fig. 7). The Carpenter Ranch maxilla belongs to a continuum of species whose members progressively increase in size, beginning in the late Arikareean with Phlaocyon annectens (Agate National Monument, Quarry 3 carnivore dens, Peterson, 1907: 53); continuing through Phlaocyon leucosteus from the lower Runningwater Formation; and terminating in the large, robust species, P. marslandensis from the Hemingford Quarries of the upper Runningwater Formation, Nebraska. Table 7 presents measurements of the upper carnassial and molars of these species taken from individual specimens, supplemented by measurements recently provided by Wang et al. (1999). Inspection of these data indicates that UNSM 44822 is best referred to Phlaocyon leucosteus of early Hemingfordian age, coeval with faunas from the lower part of the Runningwater Formation of Nebraska.

BIOSTRATIGRAPHIC SUMMARY: Taken together, the age­diagnostic genera from the Carpenter Ranch sands and gravels indicate an early Miocene age (table 8). The stage of evolution of the species, however, is earliest Hemingfordian, age­equivalent or possibly slightly older than the mammalian fauna from the lower part of the Runningwater Formation northeast of Agate post office (Northeast of Agate local fauna, MacFadden and Hunt, 1998).

THE ‘‘ARIKAREE’’ CONGLOMERATES OF N. H. DARTON (1899)

The fluvial gravels and sands here referred to the Carpenter Ranch Formation and the previously established Spoon Butte beds were recognized in the earliest geologic reconnaissance studies in southeastern Wyoming and western Nebraska conducted by the U.S. Geological Survey ( Darton, 1899; Adams, 1902). A summary of this work on the butte caprocks will supply historical background necessary to understanding UNSM field efforts in the region.

In N.H. Darton’s (1899) classic study of the geology of western Nebraska, the Ogallala and Arikaree Formations were identified, named, and mapped for the first time, and conspicuous conglomerate beds were recognized in both formations. Darton noted that ‘‘the basal Ogallala beds are mainly conglomeratic’’ and illustrated this with measured sections in the southern Nebraska panhandle ( Darton, 1899: 744, pls. 87–88). Ogallala conglomerates were recognized as crystalline gravels derived from the Rocky Mountains to the west. In contrast, the Arikaree conglomerates were lithically heterogeneous, and were restricted to the Wildcat Ridge, south of Scottsbluff, Nebraska, and the vicinity of Spoon Butte in Wyoming ( Darton, 1899: 735).

At Wildcat Ridge, Darton did not consider the conglomerates as the earliest Arikaree deposits, reporting that they occurred above the base of the Arikaree within the formation. The Arikaree conglomerates at Wildcat Ridge are composed primarily of streamrounded sandstone concretions cemented by calcite, mixed with granitic sand and gravel. Based on faunal content, they were thought to represent the channel deposits of a late Arikareean fluvial system later mapped by Vondra ( Vondra et al., 1969), who referred them to the Harrison Formation. Darton (1899: 746, fig. 212) illustrated the stratigraphic relationships of these Arikaree conglomerates in sufficient detail that they can be readily identified today south of Gering, Nebraska, along the Wildcat Ridge, where they contain the UNSM late Arikareean mammal locality Sf­105.

Darton (1899: 747) reported that the ‘‘Arikaree’’ conglomerates in the vicinity of Spoon Butte differed from those along the Wildcat Ridge in their chiefly crystalline gravel composition. We can be certain of the location and identity of these conglomerates because he published a stratigraphic profile from Spoon Butte south to Sturdivant Gap (redrafted as fig. 5), showing the conglomerates overlying Arikaree beds on the north and White River Brule clay to the south ( Darton, 1899: fig. 213, pl. 85[1]). His profile shows the conglomerates as descending in elevation southward, deposited on a ‘‘steeply sloping sur­

TABLE 6 Dental Dimensions (in mm) of the Early Miocene Amphicyonid Daphoenodon from Western Nebraska TABLE 7 Dental Dimensions (in mm) of the Borophagine Canid Phlaocyon face’’, which led Darton to regard them as strata within a single paleovalley of considerable width. He was aware that nowhere in the Spoon Butte area were these capping conglomerates overlain by Arikaree sediments (as they were at Wildcat Ridge), but this did not seem to influence his belief that they belonged to the Arikaree Formation.

George Adams (1902) extended Darton’s geological mapping into southeastern Wyoming, summarized in U.S. Geological Survey Water­Supply Paper No.70 on the geology and water resources of the Patrick and Goshen Hole quadrangles. That region at the time of Adams’ work was undergoing accelerated development for the raising of livestock, and ranching operations were rapidly invading semiarid grasslands distant from the local riv­ ers and streams. The ability to drill deep wells far from these streams, permitting the use of windmills and irrigation, led to increasing human settlement and a growing interest in a geologic survey of water resources.

Although Darton (1899) mentioned Spoon Butte and the adjacent buttes to the south, Adams (1902: pl. 3) was the first to map and clearly illustrate the topographic trends of the northern and southern buttes. He published the first and, to date, the only photograph of the Carpenter Ranch beds, shown capping Deahl Butte ( Adams, 1902: pl. VIa, Point of Rocks). Adams confirmed the ‘‘Arikaree’’ conglomerates as fluvial deposits, and he explained their occurrence at various topographic levels as explicit evidence of stream aggradation, in which ‘‘the sediments en­

TABLE 8 Mammalian Fauna of the Carpenter Ranch Formation , Goshen County, Wyoming, and Sioux County , Nebraska croached upon the higher levels’’. His chosen example was taken from Sturdivant Gap north to Spoon Butte ( Adams, 1902: 17) :

The relation of these [conglomeratic] lenses is best seen in the locality of Spoon Butte, in the Patrick quadrangle, and for a short distance to the south. In the escarpment and lone hills north of Sturdivant’s ranch, conglomerates [Carpenter Ranch Formation] are seen resting on Brule clay, or separated from it by a very thin bed of Arikaree sand. In Spoon Butte the lenses are coarse sandstones exhibiting a crossbedded structure, and their occurrence is 400 ft above the lowest exposed portion of the Arikaree. In the intermediate localities the lenses occur at somewhat lower horizons and seem to justify perfectly the interpretation [of progressively aggrading fluvial deposits] here placed upon them.

If Adams had known of the fossil content of these conglomerates capping the buttes, he might have realized that separate fluvial systems were indicated, but paleontologists who visited the buttes at about the same time failed to note any fossil occurrences of significance.

O.A. Peterson (1907: 23) of the Carnegie Museum of Natural History (Pittsburgh) was the first paleontologist to report an attempt to find fossils in the caprock of Spoon Butte. He described the caprock as ‘‘a hard pinkishgray sandstone, 35–50 ft in thickness. In this hard cap of sandstone, which is regarded as the top of the Upper Harrison beds, our party found no fossils’’. Peterson failed to find fossil mammals because the caprock itself is almost entirely barren of mammalian remains; only in the basal sands and gravels of the Spoon Butte beds at a few select localities do any mammalian fossils occur, and Peterson must have bypassed these sites. His allocation of the caprock to the Upper Harrison beds, today considered a unit of the Arikaree Group, suggests that he was aware of Adams’ description of the Spoon Butte caprock as an Arikaree conglomerate.

Later, Peterson (1909: 74–75, fig. 27) reconsidered this opinion, suggesting that the age of the caprock might be as young as Pliocene, and again remarked on the absence of fossils. It was at this time that he named the caprock at Spoon Butte the ‘‘Spoon Butte beds’’. At about the same time, the paleontologist H.F. Osborn (1909: 73, fig. 14), in a survey of Cenozoic mammal horizons of western North America, included the caprock at Spoon Butte in the broadly construed Arikaree Formation of Darton, but apparently he was unaware of Peterson’s suggestion of a much younger age.

The only published claim of anyone having found fossils at Spoon Butte occurred in 1956 when Paul O. McGrew of the University of Wyoming (in L. McGrew, 1963; also Rapp et al., 1957) collected ‘‘waterworn fragments of Pliocene(?) horse teeth... found in gravels on top of Spoon Butte and other hills in T. 27N., R.60W’’. This statement appears to be the basis for later assignment of the Spoon Butte caprock to the Pliocene or late Miocene. A loose lag gravel does occur at a few localities on top of Spoon Butte, and it contains stream­rounded and abrasion­resistant Miocene mammal fossils, including fragments of equid teeth. However, this lag gravel is a much younger deposit than the basal Spoon Butte gravel, and it contains more diverse and exotic clasts, particularly acid volcanics (rhyolite) never found in the Spoon Butte sandstone caprock or in the basal Spoon Butte gravel.

In fact, the lag gravel on Spoon Butte has a complex history and is similar in composition to loose fluvial lag gravels mantling low hills, ridges, and stream terraces at much lower topographic elevations in the valleys adjacent to Spoon Butte (fig. 3, Quaternary terraces). These granitic gravels contain conspicuous amounts of both acid volcanics and chert in a matrix of loose quartz sand (table 1, Oberg gravels). Angular blocks of the Spoon Butte caprock contained in some of these gravels indicate that they postdate the Spoon Butte beds, a fact confirmed by teeth and bone fragments of late Miocene mammals in the gravel. On the George Oberg ranch, immediately east of the butte, a high terrace gravel deposit of this type (figs. 2–4, Oberg Quarries) yielded a single tooth of Pleistocene Equus , found with water­worn bones of Clarendonian mammals, indicating that late Miocene beds were once present in the region and were subsequently removed by erosion. These acid volcanic–bearing gravels derive from mid­ and late Miocene deposits once present on the Hartville tableland that were reworked during the Pleistocene and Holocene into stream terraces and low hills that correspond to the present local topography.

The Spoon Butte caprock continued to be mistakenly regarded as Pliocene or late Miocene in age for the next two decades. Denson (1974) mapped the caprock of Spoon Butte as Ogallala Formation, at that time considered to be Pliocene or upper Miocene ‘‘poorly cemented calcareous claystone, siltstone, sandstone, and conglomerate of fluviatile origin’’. Whereas this quote adequately describes the lithologic character of typical Ogallala strata to the south, the Spoon Butte beds are characterized by silica­cemented sandstones and basal granitic/chert gravel and are thus lithically distinct from the Ogallala strata; evidently the Spoon Butte beds were not actually visited during the mapping. Crist (1977), in a hydrologic study of the area, followed Denson in showing the Spoon Butte caprock as Ogallala Formation of late Miocene age.

Love et al. (1980), in their compilation of the geology of the Torrington Quadrangle in southeastern Wyoming, indicated that the capping sandstone of Spoon Butte and the adjacent buttes of the northern butte line, the caprocks of the southern butte line (Carpenter Ranch beds), as well as similar sandstone butte caprocks east of and within the Hartville Uplift (Flattop, Sheep Mountain) were all of late Miocene age. They noted Denson’s mapping of some of these as Ogallala Formation. No mention was made of the uniform, silica­cemented sandstone lithologic composition typical of these butte caprocks. It was only when mammalian fossils from Spoon Butte and the adjacent buttes to the south became available to UNSM paleontologists from 1977–1980 that their divergent ages became apparent.

Today we recognize that the northern caprocks on Spoon Butte and adjacent buttes of the northern trend are Barstovian (mid­Miocene) in age, and the caprocks of the southern buttes are early Hemingfordian (early Miocene). These areas represent two unrelated paleovalley systems separated by a significant interval of time. Late Miocene and younger sediments were deposited in the region east of the Hartville Uplift, and then nearly entirely removed by later erosion, so that the geographic distribution of these strata is difficult to reconstruct.

We now realize that silica­cemented capping sandstones of the numerous buttes east of, and within, the Hartville Uplift are of diverse ages. Nevertheless, insofar as is known, all fall within the Miocene. These caprocks record multiple episodic intervals of fluvial deposition, during which crystalline sands and gravels derived from the Laramie Mountains and Hartville Uplift were transported eastward to the bordering plains.

The provenance of the Carpenter Ranch gravels is of particular interest because, together with the gravels of the Runningwater Formation, the clast types indicate derivation from either the Precambrian core of the north­central Laramie Mountains or possibly from the southern Hartville Uplift. Archean granitic and gneissic rocks cut by numerous amphibolite dikes form the greater part of the northern and central Laramie Mountains ( Condie, 1969; Sims and Day, 1999; Sims and Stein, 2003). To the south, these Archean rocks are bordered by the Proterozoic Laramie anorthosite body.

Archean granites of the northern Laramie Mountains are separated from an orthogneiss terrain to the south by the Laramie Peak shear zone ( Condie, 1969; Chamberlain et al, 1993; Resor 1996; Sims and Day, 1999), also termed the Garret­Fletcher Park­Cottonwood Park shear zone ( Snyder et al., 1995), which includes deformed granitic and gneissic rocks, quartzite, and amphibolite dikes. Mafic amphibolite dikes of the north­central Laramie Mountains often exhibit a relict igneous texture that characterizes Carpenter Ranch amphibolite clasts. Streams accessing the Precambrian core of the north­central Laramie Mountains are capable of transporting granitic/gneissic gravel with high amphibolite content. Sampling of surficial stream and pediment gravels east of the Laramie Mountains from the vicinity of Wheatland north to Esterbrook suggests that terrace gravel of the Laramie River north of Wheatland carries a granitic/amphibolite/anorthosite clast content most similar to that of the Carpenter Ranch Formation.

Mapping of Precambrian rocks of the southern Hartville Uplift ( Millgate, 1965; Snyder, 1980; Sims et al., 1997; Sims and Day, 1999) identified a granite dome (Haystack Range) in contact with quartzofeldspathic schists, metadiabase/amphibolite dikes, and pegmatite east of the north­trending Hartville Fault. Metadolomite, metaconglomerate, schist, phyllite, maroon quartzite, and banded ironstone of the Precambrian Whalen Group occur west of the fault. Granite, gneiss and amphibolite in the Haystack Range are similar to Carpenter Ranch gravel, which lacks metasediments, despite proximity of the Whalen Group only 20 miles (; 32 km) west of the Carpenter Ranch Formation. Arikaree sediments buried much of the Hartville Uplift in the early Miocene and it remains an open question as to what extent the southern part of the uplift was accessible to these streams.

Carpenter Ranch and Runningwater gravels are granitic in composition, with amphibolite (metabasalt, metagabbro) common, gneiss, chert and quartzite present, and anorthosite rare. Based on clast content, the provenance for these early Miocene gravels would seem to be the north­central Laramie Mountains. The Haystack Range possibly contributed during the Carpenter Ranch fluvial cycle, but lack of metasediments in the gravels favors the Laramie Mountains as the principal source. Scarcity of anorthosite clasts suggests that the streams originated in granitic terrain north of the Laramie anorthosite body.

The west­to­east stream­flow trend of the Carpenter Ranch Formation caprock belt is evident from topographic maps and satellite photoimages. Our field inspection of Carpenter Ranch outcrops from Duncan Buttes in Wyoming eastward to Carpenter Butte in Nebraska revealed a major paleochannel incised into the White River beds (fig. 24). It represents the deepest incision of the paleovalley and confirmed the west­to­east alignment of the paleotrend. The paleovalley then turns to the northeast upon entering Nebraska .

The easternmost exposures of the Carpenter Ranch Formation that we have identified occur at the northeastern limit of East Sturdivant Butte in the NE1/4, NE1/4, sec. 27, T27N, R57W (Agate SW 7.5­minute quadrangle). These outcrops indicate that the paleovalley has reoriented to a northeasterly course. At this locality, Carpenter Ranch outcrops disappear beneath Quaternary dune sands, which then blanket the landscape for; 10 miles to the northeast. When Miocene sediments eventually reappear beyond the dune field in the eastern half of T28N, R56W, granitic gravel and tuffaceous sandstones have been assigned to the Runningwater Formation and represent the westernmost outcrops of that rock unit currently recognized ( Skinner et al., 1977). The mammals from tuffaceous sandstones of the lower Runningwater beds northeast of Agate are close in age to the Carpenter Ranch fauna and, together with the paleovalley trend, suggest that the Runningwater paleovalley identified by Skinner et al. (1977) is a direct continuation of the Carpenter Ranch trend (fig. 25).

The Carpenter Ranch mammals appear less advanced than those from the upper Runningwater granitic gravels and sands that yielded most of the classic Runningwater fauna to the east in Dawes and Box Butte counties, where most productive quarries have been located. However, the stratigraphically lowest sites in that area, such as Runningwater Quarry, show an affinity to the Carpenter Ranch fauna and to the fauna from lower Runningwater strata northeast of Agate. The evident similarity in age of the Northeast of Agate fauna to the Carpenter Ranch mammals suggests that the pale orange­brown tuffaceous sandstones northeast of Agate are a fine­grained sedimentary equivalent of the Carpenter Ranch fluvial sands and gravels, the two lithofacies togeth­ er representing the earliest deposits of the Runningwater/Carpenter Ranch paleovalley system established in the earliest Hemingfordian of the central Great Plains.

POST­LARAMIDE EVOLUTION OF THE ROCKY MOUNTAINS

Coarse epiclastic debris dispersed by post­ Laramide fluvial systems into the mid­con­

tinent documents a mid­ to late Cenozoic record of episodic uplift of the adjacent Rocky Mountain ranges. Mammalian fossils preserved in the sedimentary fill of major paleovalleys emerging from the uplifts in southeastern Wyoming, northeastern Colorado, and western Nebraska provide the basis for a highly resolved tectonic chronology. These paleovalleys are often identified by a reversed topography of geographically aligned butte caprocks, forming conspicuous linear trends within and east of the uplifts. Recovery of mammalian faunas from the Carpenter Ranch and Spoon Butte paleovalleys in southeastern Wyoming demonstrated unexpected ages for two such fluvial systems. Similar fluvial trends east of the Laramie Mountains and Hartville Uplift could also represent more diverse ages than previously supposed. Many remain to be examined for biochronologically useful faunas .

The oldest widely recognized post­Laramide fluvial activity east of the Laramie Mountains begins in the late Eocene with widespread sands and gravel of the White River Group ( Stanley, 1976). Following this initial pulse of coarse granitic debris from the Laramie Mountains, transported eastward by basal White River channels of Chadronian age, no significant evidence of regional uplift recurs until the initiation of late Oligocene stream incision, evidenced by basal Arikaree paleovalleys flowing eastward from the Laramie Mountains and Hartville Uplift ( Vondra et al., 1969; Swinehart et al., 1985). Early Arikareean mammals and 40 Ar/ 39 Ar radiometry (;28.1–28.3 Ma) date this tectonic event ( Tedford et al., 1996). Removal of a considerable volume of lower Arikaree Group strata by late Arikareean incision is documented by east­northeast trending paleovalleys in western Nebraska at Wildcat Ridge ( Vondra et al., 1969) and along the Niobrara River at Agate National Monument ( Hunt, 1985b). At the latter locality both mammals and an 40 Ar/ 39 Ar date of 22.9 6 0.08 Ma ( Izett and Obradovich, 2001) establish the approximate time of regional incision, and deposition of the Harrison Formation.

Fine­grained sheetlike sediments of the Arikaree Group continued to accumulate east of the Laramie and Hartville uplifts, represented by the Anderson Ranch Formation ( Hunt, 2002). These sediments incorporate a significant eolian tuffaceous component with latest Arikareean mammals. Only a single major paleovalley (fig. 3, Lay Ranch beds) within the Anderson Ranch Formation has been recognized and mapped; it deeply incises undifferentiated Arikaree rocks at Spoon Butte in southeastern Wyoming ( Hunt, 1985a, 1990). This valley, emerging from the vicinity of the southern Hartville Uplift, carried fine­grained tuffaceous fluvial sands and silts with only minor amounts of epiclastic granitic sand and granule gravel. The scarcity of latest Arikareean paleovalleys on the Hartville tableland suggests that this was a tectonically quiescent interval. The granitic cores of the mountains to the west were likely nearly buried by tuffaceous Arikaree strata, evidenced today by the onlap of the Anderson Ranch Formation on the granites of the Hartville Uplift ( Hunt, 2002: fig. 1) and by residual Arikaree deposits bordering the Laramie Mountains.

The conclusion of Arikaree deposition east of the Hartville Uplift at;19 Ma is evidenced by the development of multiple siliceous paleosols found throughout the upper part of the Anderson Ranch Formation (previously Upper Harrison beds of Hunt, 1990: figs. 7, 8, 10). Densely packed, fine siliceous rhizoliths of uniform diameter (0.5 mm) form highly anastomosed root networks in these paleosols typical of grasslands. The rhizolith networks are restricted to level geomorphic surfaces that can be traced over 1200 km 2 east of the Hartville Uplift, demonstrating the regional extent of these early Miocene grasslands. However, Arikaree deposition continued into Hemingfordian time in central Wyoming (Split Rock Formation, Love, 1970; Munthe, 1988) and eastward in the vicinity of the Laramie Mountains ( McGrew, 1963). An 40 Ar/ 39 Ar date (17.4 6 0.08 Ma, Izett and Obradovich, 2001) and mammal fauna from the Split Rock Arikaree outcrops of Munthe (1988) indicate a medial Hemingfordian age.

Evidence of a major early Miocene tectonic event involving regional uplift of the Laramie Mountains and Rocky Mountain Front Range is documented by the appearance of paleovalley systems of early Hemingfordian age (fig. 26) carrying coarse granitic debris into the central Great Plains ( Skinner et al., 1977; Hunt, 1985a; Flanagan and Montagne, 1993; Tedford, 1999). Trending east­northeast from the mountains, a northern paleovalley transporting sediments of the Carpenter Ranch and Runningwater Formations traverses southeastern Wyoming and continues into western Nebraska. The gravel composition of this system is chiefly granitic, lacking any evidence of rhyolitic or other acid­volcanic clasts ( Yatkola, 1978; Hunt, 1985a). To the south in northern Colorado, a second major system represented by sediments of the Martin Canyon Formation ( Scott, 1982; Tedford, 1999) follows a similar direction across northeastern Colorado. Here the lithic composition includes not only granitic debris, but also volcanic clasts (rhyolite, andesite), indicating a source in the Rocky Mountains involving the Never Summer Range–Specimen Mountain volcanics ( Wahlstrom, 1944; Corbett, 1968).

These paleovalleys contain earliest Hemingfordian mammal faunas that firmly establish their contemporaneity and allow the sediments to be correlated with considerable precision. The occurrence of the large oreodont Merycochoerus magnus matches them with a third locale in the Rocky Mountains at Middle Park, Colorado, where early Miocene sediments of the Troublesome Formation overlie a weathered Cretaceous paleosurface ( Izett, 1968). Evidently, regional uplift of the Front Range–Laramie Mountain axis initiated simultaneous deposition on both the western and eastern flanks of the range at;18.2–18.8 Ma. It has not escaped our notice that the early Hemingfordian paleovalleys of the High Plains trend northeast, indicating a topographic gradient in that direction. The sediments of these early Miocene paleovalleys, primarily confined to the linear trend of the stream courses, differ in the lithologic composition of their axial gravels, which clearly reflect local source terrain in the mountains to the west. The early Hemingfordian valley fills are followed in time by mid­ and late Miocene Ogallala sands and gravels within paleovalleys and alluvial aprons derived from continuing regional uplift of the Rocky Mountain plateau.

Figures

Fig. 23. Left maxilla of the canid Phlaocyon leucosteus (UNSM 44822) with P3­M2, Cow Trail Notch local fauna, Carpenter Ranch Formation, East Sturdivant Butte, Sioux County, Nebraska. Stereopair. Scale bar, 1 cm.