The Screw Bean Local Fauna is the earliest Hemphillian fauna of the southwestern United States. The fossil remains occur in all parts of the informal Banta Shut-in formation, nowhere very fossiliferous. The formation is informally subdivided on the basis of stepwise fining and slowing deposition into Lower (least fossiliferous), Middle, and Red clay members, succeeded by the valley-filling, Bench member (most fossiliferous).
Identified Carnivora include: cf. Pseudaelurus sp. and cf. Nimravides catocopis, medium and large extinct cats; Epicyon haydeni, large borophagine dog; Vulpes sp., small fox; cf. Eucyon sp., extinct primitive canine; Buisnictis chisoensis, n. sp., extinct skunk; and Martes sp., marten. B. chisoensis may be allied with Spilogale on the basis of mastoid specialization. Some of the Screw Bean taxa are late survivors of the Clarendonian Chronofauna, which extended through most or all of the early Hemphillian. The early early Hemphillian, late Miocene age attributed to the fauna is based on the Screw Bean assemblage postdating oreodont and predating North American edentate occurrences, on lack of defining Hemphillian taxa, and on stage of evolution.
INTRODUCTION
Naming and Importance of the Screw Bean Local Fauna:
The name “Screw Bean Local Fauna,” Banta Shut-in formation, Trans-Pecos Texas (fig. 9.1), was introduced by Stevens and Stevens (1989: 77). The name, misprinted as Screw “Beam” in Janis et al. (1998b: 636), comes from “tornillo,” Spanish for screw, or, in local context, from the tightly coiled seedpods of the screw-bean mesquite, common along some local arroyos. To date, the Screw Bean Local Fauna contains 4 taxa of lower and 18 taxa of higher vertebrates. Considered together, the fossils indicate a late Miocene age (early early Hemphillian) that is not widely found in southwestern North America, and fill a paleobiogeographic gap. In Trans-Pecos Texas and adjacent Chihuahua and Coahuila, Mexico, they provide an age determination for postvolcanic (<18–20 Ma; Henry et al., 1998) rocks (Banta Shut-in formation; Stevens and Stevens, 1989) that contain them. Rocks of the interval between the earliest Miocene (bolsons, the Big Bend area; Stevens et al., 1969; Stevens, 1977), and the late Pliocene–early Pleistocene (Blancan bolson fills, western Trans-Pecos Texas; Strain, 1966, 1980; Akersten, 1972) were previously unknown in the region. The tectonic importance of the age and body of rock is shown in discussions of the Texas Lineament, the eastern Basin and Range Province, and the Rio Grande Rift and related areas (e.g., Dickerson and Muehlberger, 1994; Henry, 1998; Stewart, 1998; Henry and Aranda-Gomez, 2000). This paper reports the fossil carnivores from these rocks, and presents a local informal stratigraphy sufficient to place them in stratigraphic position and in depositional context.
Discovery and Disposition of Fossils:
The first late Miocene fossils found in Trans-Pecos Texas were discovered in 1980 by Jim Liles, Chief Ranger, Big Bend National Park. Mr. Liles reported the fossils to Dr. Wann Langston, Jr., then Director of the Vertebrate Paleontology Laboratory, The University of Texas at Austin. Dr. Langston was helicoptered into the initial fossil site, TMM 42247 (Liles measured section; fig. 9.2F), collected more specimens, and documented its location. We have collected more fossils since 1980. Supraspecific classification used in this paper follows that suggested by McKenna and Bell (1997).
All fossils and rock samples were collected in accordance with regulations outlined by collecting permits, 1980–1997, issued to the Vertebrate Paleontology Laboratory, Balcones Research Center, The University of Texas at Austin, by the United States Department of the Interior. Other permits were issued directly to us by Science and Resources Management, Big Bend National Park, 1998–2000. All specimens are preserved in the TMM collections; measurements in millimeters. This paper is a contribution of the Vertebrate Paleontology Laboratory, The University of Texas at Austin. Systematic descriptions are the responsibility of MSS, and JBS is responsible for discussion of stratigraphy and geologic setting.
Institutional Abbreviations
AMNH American Museum of Natural History, New York
DMNH Denver Museum of Natural History, Denver
F:AM Frick Collection, AMNH, New York
KUMNH University of Kansas Museum of Natural History, Lawrence
KUMVP University of Kansas Museum of Vertebrate Paleontology, Lawrence
LACM Los Angeles County Museum, Los Angeles
MU Midwestern University Department of Mammalogy, Wichita Falls
MUC:FV Midwestern University Collection, Fossil Vertebrates, Wichita Falls
TMM Texas Memorial Museum, The University of Texas at Austin
UF University of Florida, Gainesville
UO University of Oregon, Eugene
UCMP University of California Museum of Paleontology, Berkeley
UMMP University of Michigan Museum of Paleontology, Ann Arbor
UOMP University of Oklahoma Museum of Paleontology, Norman
USNM United States National Museum, Smithsonian Institution, Washington, D.C.
WTM West Texas Museum, West Texas State University, Canyon
GEOLOGIC SETTING AND STRATIGRAPHY
Previous Work:
Neogene sediments of Trans-Pecos Texas (fig. 9.1) have long been associated with basins formed by extensional faulting (Hill, 1900; Tolman, 1909; Groat, 1972). Udden (1907) was the first to call attention to sediments of this kind in the Dugout Wells–Lower Tornillo Creek area (figs. 9.3, 9.4), Big Bend National Park, Texas. He noted the conspicuous twofold lithology of the deposits, and called them the Dugout clays and gravels. Placement of the roads (now abandoned) that Udden traveled make probable the identification of the “clays” as the rocks we call the Banta Shut-in formation. The “gravels” probably refers to coarse deposits of a considerably younger system of composite alluvial fans (informally, Estufa Canyon formation). Maxwell et al. (1967: pl. II, 152–153) mapped the Dugout clays and gravels as Quaternary-Tertiary older gravels, but correlated the “gravels” portion with undoubted Pleistocene deposits near Grapevine Springs. Thurwachter (1984a, 1984b, 1984c) studied the sedimentology of these “older gravels.” He viewed the “clays” and “gravels” as the result of two depositional, basin-filling events, and, in part, mapped them separately. Because his mapping units were established on the basis of depositional styles that recurred at various times, he included late Miocene, ?Blancan, and late Pleistocene deposits in the same unit (“facies,” or “member”) in some places. Thurwachter (1984c) called the “clays,” “La Noria member,” and the gravels “Estufa member,” but did not assign a formation name. His “members” do not entirely coincide with any units we recognize.
Tectonic and Structural Relationships of Late Miocene Rocks:
King (1935: 251–261) considered the initiation of normal faulting and associated basin filling to be “later Tertiary (Miocene or Pliocene),” noted that movement on the faults continued into recent times, and commented on the absence of a major through-flowing stream during basin formation and filling and the later integration of the Rio Grande system. These points stand refined today, but only slightly modified (e.g., Muehlberger et al., 1978; Stevens and Stevens, 1985, 1989, 1990; Xie, 1998; Henry et al., 1998). Tectonic and structural geology of the area has been the subject of a number of studies, the most pertinent of which are those of Muehlberger (1989), Dickerson and Muehlberger (1994), and Dickerson (1995), who defined and discussed the Tornillo graben, in which later Miocene rocks are preserved. Tornillo graben (figs. 9.3, 9.4; west northwesterly elongated; ≈32 × 11 km) lies between the Chisos Mountains and the Sierra Del Carmen in the eastern part of Big Bend National Park, southern Brewster County, Trans-Pecos Texas. It is bounded on the west-southwest by Dugout Wells fault zone (figs. 9.3B, 9.4; Stevens and Stevens, 1986, 1989, 1990), on the north by an area of persistent uplift (during at least the Cenozoic; Dickerson, 1980, 1995; Stevens and Stevens, 1990), and on the east-northeast by Sierra Del Carmen (Moustafa, 1988). The southern margin is not well understood, but is a structural high, possibly related to the tentatively proposed Santa Elena zone (Dickerson, 1980). Tornillo graben is within, but near the northeastern margin of, tectonic provinces mentioned above. In addition, the Sierra Del Carmen is commonly discussed as the Laramide front in Trans-Pecos Texas (e.g., King, 1935; Muehlberger, 1989).
Extent, Structural, and Stratigraphic Relationships of Later Miocene Rocks:
The later Miocene rocks in the Tornillo graben usually rest on rocks near the top of the Late Cretaceous Aguja Formation, but as young as Eocene along the Dugout Wells fault zone. To the east of Lower Tornillo Creek from near C′ to near G′ (figs. 9.3B, 9.4), later Miocene rocks rest with substantial angularity on the San Vicente Member, Boquillas Formation or on lower parts of the Pen Formation. Reddish to reddish buff Banta Shut-in rocks are most deeply dissected and widely exposed to the west of Lower Tornillo Creek, between the creek and Park Route 12 (Panther Junction–Rio Grande Village road; fig. 9.3B). Remnants of pediment gravels and pre-pedimentation alluvial fan deposits cover the Banta Shut-in formation adjacent to Route 12 (fig. 9.3B). Good exposures of the Banta Shut-in formation southwest of Route 12, somewhat obscured by pediment cover, inadvertently had been mapped as Aguja Formation (Maxwell et al., 1967: pl. II). The irregularly northwest-tapering outcrop area of the Banta Shut-in formation, apparently a remnant of a more extensive deposit, is about 200 km2 (figs. 9.3, 9.4), but an additional ≈100 km2 of the formation is covered by the possibly Blancan Estufa Canyon formation. The exposure area of the units taken together has the form of an elongate oval flattened against the Dugout Wells fault zone (figs. 9.3B, 9.4; Stevens and Stevens, 1989), which suggests the sag common along high-angle normal faults (Jackson and McKenzie, 1983; Alexander and Leeder, 1987; Leeder and Gawthorpe, 1987; Schlische, 1991). Internal structures, high-angle normal faults, and a westerly-facing monocline appear consistent with structures discussed for extensional basins by Schlische (1995) if associated with down-to-the-east displacement on the Dugout Wells fault zone.
Banta Shut-in Formation and Association of Fossils with Deposits:
The later Miocene rocks were deposited under various regimes that produced distinctive lithologies. Differences in depositional architecture, coarse clast lithology, grain size, and mud content suggest informal division into a Lower (symbol: Tbslm; thickness: ≈60–105 m), Middle (Tbsmm: ≈75–85 m), Red clay (Tbsrc: >80 m), and a Bench member (Tbsbm; >90 m; figs. 9.2, 9.3B, 9.4). Contacts between the Lower and Middle, and between the Middle and Red clay members are gradational over a thickness of a few meters, and drawn arbitrarily at soil zones. The Bench member rests erosionally, but not apparently angularly, on Lower and Middle member rocks to the north of Liles locality, and is inferred to have an erosional contact with the Red clay member south and west of Liles locality (figs. 9.2–9.4). There is no evidence to suggest hydrologic closure of the basin during deposition. Banta Shut-in formation does not record much penecontemporaneous tectonic activity, beyond that necessary to form and maintain a depositional basin, and that necessary to establish an initially rugged source area for the coarse Lower member.
Fossils have been found wherever the deposits are well exposed, except in the coarsest conglomerates low in the Lower member. Most fossils come from muddier parts of the Middle and Bench members, but isolated teeth and tortoise “rings” have been found low in the Red clay member. Rapidly eroding footprints, including a coyote-size canid, have been found in recently fallen blocks from a probable floodout waterhole (Melville and Erskine, 1986; Tooth, 1999) deposit in the Cam sand (Middle member; figs. 9.2, 9.4), near the Hellhole measured section (fig. 9.3A, section B). Only three localities have proved worthy of repeated prospecting. Of these, two moderately productive Middle member localities include one near the base of the Tortoise Leg measured section (at ≈8 m; figs. 9.2E, 9.3A, section E; TMM 42864), and another at the north base of Hill 2362 (“Comanche ridge”; TMM 42124; fig. 9.2D). The most productive site is the area of the Liles locality section, benches 2–3 (figs. 9.2F, 9.3A, section F; TMM 42247, 42252), Bench member (Roy's Peak Quadrangle, 1971).
SYSTEMATIC PALEONTOLOGY
ORDER CARNIVORA BOWDICH, 1821
SUBORDER FELIFORMIA KRETZOI, 1945
FAMILY FELIDAE FISCHER DE WALDHEIM, 1817
SUBFAMILY FELINAE FISCHER DE WALDHEIM, 1817
Pseudaelurus Gervais 1850 ?Pseudaelurus sp. Figure 9.5A–T; table 9.1
Referred Material:
TMM 42247-237, proximal end, left ulna; TMM 42247-144, distal end, right humerus; TMM 42247-136, distal end left humerus; TMM 42247-127, proximal end, left radius; TMM 42247-114, distal end of metacarpal or metatarsal; and TMM 42247-182a and -182b, proximal and distal ends of a metacarpal III.
Stratigraphic Occurrence:
TMM locality 42247, from float derived from the top of Bench 2, Liles measured section (fig. 9.2F), Bench member, Banta Shut-in formation, Big Bend National Park, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description:
The presence of ?Pseudaelurus in the Screw Bean Local Fauna rests on the identification of postcranial fragments. The proximal end of an ulna, TMM 42247-237 (fig. 9.5N, O), lacks the epiphysis and is broken across the coronoid process. The bone has a strong, squared olecranon process and a well-splayed semilunar notch. It is similar to Puma concolor, the mountain lion, but has a shorter olecranon process. The head of the proximal radius (AP and T diameters, 165 and 25.1 mm) has a shallow capitulum with an elevated, pointed posterior margin, and is similar to P. concolor (fig. 9.5L, M). The two distal humeral fragments are referred to Pseudaelurus because of size and felid morphology (fig. 9.5P–T). Only TMM 42247-144 preserves part of the entepicondylar foramen (fig. 9.5P, Q). The condyle lacks a supratrochlear foramen. The humeral condyles are similar in size and morphology to those of humeri of P. concolor.
The distal ends of two metacarpal or metatarsal elements, TMM 42247-182b, 42247-114 (fig. 9.5A–G) also represent this puma-sized felid. Their distal articulations are well rounded, have an extensive radius of curvature, strong ligament attachments, and great similarity with homologues in the extant Puma concolor. The proximal end, TMM 42247-182a, of the left third metacarpal, identical in structure to P. concolor, has been recovered (fig. 9.5H–K).
Discussion:
The Banta, scavenger-ravage, Puma-sized, felid limb bone and toe fragments have been difficult to identify because no comprehensive modern revision of Neogene North American cats has been published. The bones indicate a cat slightly less than one half the size of cf. Nimravides catocopis, discussed below. Hibbard (1934) described a Puma-sized, saber-toothed kind of felid from Edson Quarry, late Hemphillian, Sherman County, Kansas, as Adelphailurus kansensis. Harrison (1983) was the first to describe some of the postcranial bones of this animal. The distal end of the Screw Bean humerus is transversely less expanded, and has a longer entepicondylar foramen than in Adelphailurus kansensis. Also, the proximal end of the Screw Bean ulna is deeper and more distinctly squared than in Adelphailurus, and more like Puma. Although generally size similar to Puma concolor, A. kansensis is not ancestral to P. concolor, and the late Hemphillian age of the Edson Quarry fossils would seem to deny a faunal relationship between the Edson and the Screw Bean assemblages. The Banta Puma-sized fragments tentatively are referred to Pseudaelurus Gervais, the ancestor of the modern cats.
The genotypic species of Pseudaelurus is P. quadridentatus (Blainville) from the middle Miocene of Sansan (Gers), France. Leidy (1858: 22; see 1869: 52, and pl. I, fig. 8) was the first to recognize Pseudaelurus in North America through the description of Felis (Pseudaelurus) intrepidus, based on a lower jaw from the “sands of the Niobrara River,” Valentine Formation, later Barstovian. P. intrepidus and P. intrepidus sinclairi Matthew (1918), early Barstovian Lower Snake Creek Fauna, Olcott Formation, Sioux County, Nebraska (Skinner et al., 1977: 347), are much older and smaller than the Screw Bean felid. Pseudoaelurus marshi Thorpe (1922: 446) is based on YPM 12865, a lower jaw from near the mouth of Minnechaduza Creek, Cherry County, Nebraska, Niobrara River Fauna (Webb, 1969: 35), late Barstovian. P. marshi is smaller and considerably older than our species. MacDonald (1954: 67) described P. aeluroides, from the northeast rim of Sinclair Draw, Snake Creek Fauna, late Clarendonian (Skinner et al., 1977: 347), which also is much smaller than the Screw Bean cat.
Nimravides Kitts, 1958 cf. Nimravides catocopis (Cope) Figure 9.6A–M; table 9.2
Machairodus catocopis Cope 1887: 1019.
Nimravides catocopis (Cope), Martin and Schultz, 1975: 60.
Referred Material:
TMM 42247-33, posterior one half of m1; TMM 42247-138, partial right upper canine lacking enamel; TMM 42124-21, distal one half of left humerus; and TMM 42247-171, pollex, right medial phalanx.
Stratigraphic Occurrence:
TMM 42124, Comanche ridge measured section (fig. 9.2D); and TMM locality 42247, from float derived from the top of Bench 2, Liles measured section (fig. 9.2F), Middle member and lower part, Bench Member, Banta Shut-in formation, Big Bend National Park, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description:
The posterior one half of a slightly worn, African lion-sized m1, broken through the carnassial notch (fig. 9.6A, B, arrow), is the most diagnostic specimen. The apex of the protoconid is broken away. A sharp ridge extends anterior from the protoconid to the vertical carnassial notch. A distinct metaconid is located posterior to the protoconid. A sharp, low ridge, extends posterior from the apex of the metaconid to the base of the enamel. The partial canine (fig. 9.6C, D), a little less than one half as broad transversely as long anteroposteriorly, is believed to be an upper right on the basis of its limited curvature and flattened medial surface. The weathered fragment, about the size of, but differently shaped than in the African lion lacks the tip, root, and enamel.
The distal one half of humerus TMM 42124-21 (fig. 9.6I–M) is broken just below the deltoid tuberosity. The condyle is well preserved, but the medial part of the trochlea has been deformed downward whereas the external part of the trochlea is deformed upward (arrows) as preserved (fig. 9.6K–M, restored). A large entepicondylar foramen and entepicondylar bridge, a deep olecranon fossa, and an especially enlarged medial epicondyle are developed. The deep coronoid fossa lacks a supratrochlear foramen. The humerus closely resembles that of the modern mountain lion, Puma concolor, but is much larger. A nearly complete right pollex (fig. 9.6E–H) is the same magnitude larger than homologues in P. concolor, as the other bone fragments here referred to Nimravides catocopis. The phalanx, however, differs from P. concolor in greater medial flexion, in being longer and slightly narrower laterally, and in some of the details of the distal articulation. If the toe is identified correctly, this animal had a relatively larger and more diverging pollex than the modern cats used for comparison.
Discussion:
The m1, canine, phalanx, and humeral fragment show that an African lion-sized felid with oval canine is present in the Screw Bean Local Fauna. Lion-sized, scimitar-toothed cats of comparable age are Nimravides Kitts (1958) of the late Clarendonian to the end of the early Hemphillian (Tedford et al., 1987; Martin, 1998), and Machairodus Kaup (1833), late early to latest Hemphillian (Tedford et al., 1987). Machairodus differs from Nimravides by having a serrated upper canine and an indistinct or absent m1 metaconid. Additionally, Machairodus arrived from Eurasia along with other index taxa (unknown in the Screw Bean Local Fauna), in the late part of the early early Hemphillian, during a low sea level stand. The Screw Bean Local Fauna is believed to predate that immigration event. The distinct m1 metaconid suggests that our species is Nimravides, and not Machairodus.
Martin (1998: 238) recognizes three species of Nimravides, N. catocopis (= Pseudothinobates Kitts), N. pedionomus (Macdonald, 1948b, originally Pseudaelurus), and N. galiani (Baskin, 1981). N. pedionomus (misprinted as pediomus in Martin, 1998) is based on a puma-sized jaw from the Cap Rock base of the Ash Hollow Formation, Minnechaduza Fauna, early Clarendonian, Cherry County, Nebraska. Baskin (1981: 137) retained P. pedionomus in Pseudaelurus and suggested that it is probably ancestral to Nimravides. The Screw Bean species under discussion differs from P. pedionomus by a more prominent metaconid, wider talonid, and larger size. Additionally, the Screw Bean Local Fauna is certainly younger than early Clarendonian.
Nimravides galiani (Baskin, 1981) is based on a partial jaw with a canine and m1, UF 24461, from the latest Clarendonian, Love Bone Bed Local Fauna, Alachua County, Florida. N. galiani has also been identified from the Shoreline assemblage (Leite, 1990), Lake McConaughly, Ash Hollow Formation, Keith County, Nebraska, late Clarendonian. Our carnassial fragment, TMM 42247-33, closely resembles the posterior one half of UF 24461 with its distinct metaconid, but the Texas species is slightly larger. Of the three recognized species of Nimravides, the Screw Bean fragmentary material matches N. catocopis best in size and morphology.
The earliest occurrence of Nimravides catocopis is from the Black Hawk Ranch (see Macdonald, 1948a) and the North Tejon Hills faunas, late Clarendonian (Tedford et al., 1987). N. catocopis has been reported from various early Hemphillian faunas, such as the Gabaldon Badlands, New Mexico, Arnett, Capps Pit, Port-of-Entry Pit, Higgins (Sebits's Ranch), Texas and Oklahoma, and the upper part of the Ash Hollow Formation, Lemoyne, Feltz Ranch and Oshkosh faunas, Nebraska (Kitts, 1958; Martin and Schultz, 1975; Schultz, 1977; Tedford et al., 1987; Voorhies, 1990; Lozinsky and Tedford, 1991). The youngest occurrence as known is in the Box T, Wolf Creek fauna, Lipscomb County, Texas (Schultz, 1977: 69), early Hemphillian.
The presumed late Hemphillian occurrences of “Nimravides catocopis” reported for the Coffee Ranch Fauna, Hemphill County, Texas (see Burt, 1931: 262; Dalquest, 1969), the Optima fauna, Texas County, Oklahoma (Savage, 1941: 697), the Yepómera fauna, State of Chihuahua, and the State of Guanajuato, Mexico, are based on specimens of Machairodus coloradensis Cook (Schultz, 1977; Dalquest, 1983; Lindsay et al., 1984; Carranza-Castañeda and Miller, 1996).
SUBORDER CANIFORMIA KRETZOI, 1943
INFRAORDER CYNOIDEA FLOWER, 1869
FAMILY CANIDAE FISCHER DE WALDHEIM, 1817
SUBFAMILY BOROPHAGINAE SIMPSON, 1945
Epicyon Leidy, 1858
Epicyon haydeni Leidy, 1858 Figure 9.7E–G, P, Q; table 9.3
Canis (Epicyon) haydeni Leidy 1858: 21.
Epicyon haydeni (Leidy): Baskin, 1980: 1350.
Epicyon haydeni (Leidy): Baskin, 1998b: 207.
Epicyon haydeni (Leidy): Wang et al., 1999: 252.
Bone-eating dog, unident.: Stevens and Stevens, 1989: 79.
Referred Material:
TMM 42252-28, associated, detached, right and left maxillary fragments with M1–2 and partial P4 root; and TMM 42247-86, distal end of left humerus, tentatively referred.
Stratigraphic Occurrence:
Derived from float from the top of Bench 2, and from between Bench 2 and 3, Liles measured section (fig. 9.2F), Bench member, Banta Shut-in formation, Big Bend National Park, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description:
The associated but detached maxillae fragments preserve a root for the right P4, and right and left M1–2 (fig. 9.7E–G). Both M1s have the paracone broken away. M1–2 are strong, crushing, well-worn teeth as preserved. They are considerably larger than homologues in modern gray wolves, Canis lupus Linnaeus. M1 is canid in outline but has a broad, little posteromedially flexed talon. A slight anteromedial cingulum is present on M1. The M2 is about three-fourths the size of M1.
The distal end of the humerus, TMM 42247-86, is larger than counterparts in Canis lupus used for comparisons. The distal end has the distinctly “squared” medial profile, although the medial epicondyle is broken away, and a small supracondylar foramen, typical of canids (fig. 9.8P, Q). The humeral fragment has a relatively broad troachlea, unlike the more transversely constricted articulation of the modern cursorial canids. The olecranon fossa of TMM 42247-86 is constricted less than in Canis. The presence or absence of an entepicondylar foramen cannot be determined due to incompleteness. The humeral fragment is morphologically similar to that figured by Munthe (1989) for E. “validus,” but is not as large. TMM 42247-86 is referred tentatively to Epicyon haydeni.
Discussion:
Baskin (1980, 1998a, 1998b) and Wang et al. (1999) have aptly reviewed the chaotic taxonomic history of Epicyon, and have paired cranial and other skeletal elements to the type lower jaw. Wang et al. (1999: 266) reviewed E. haydeni, early Clarendonian through the early Hemphillian. E. haydeni survived, apparently in isolation, on into the late Hemphillian in Florida. The Screw Bean specimen adds very little to our knowledge of this species, except to expand the geographic distribution of this broad-ranging taxon (Baskin, 1998b) southward toward Mexico. The identification of our specimens as E. hayden, rests on morphology and size. Examples of very worn teeth of E. haydeni (see Wang et al., 1999), F:AM 61428 and F:AM 61523B, Port-of-Entry Pit, Arnett fauna, Ellis County, Oklahoma, late early Hemphillian, closely match in size and shape TMM 42247-86, but our specimen is slightly smaller than F:AM 61512, an unworn tooth. Richard H. Tedford (written commun., Dec. 7, 1993) provided measurements of five specimens of the sympatric species, Epicyon saevus, from Port of Entry Pit. The observed range for AP M1 ranged from 18.3 to 20.5 mm, and the TM1 observed range varied between 21.6 and 24.0 mm. These dimensions are somewhat smaller than for the well-worn teeth of TMM 42252-28. Additionally, he graciously provided measurements of several specimens from Aphelops Draw, Johnson Member, Snake Creek Formation, late early Hemphillian. An APM1 diameter at 22.5 and a TM1 at about 25.3 mm were obtained for AMNH 20482, and an APM1 at 22.0 mm and a TM1 at 24.0 mm were obtained for AMNH 20482a. The Screw Bean teeth fall within the observed range but below the mean for a large statistical sample of E. haydeni (Wang et al., 1999: 389). If unworn, the Screw Bean teeth would have yielded slightly larger measurements.
Borophagine canids are believed to have been solitary hunters as determined from brain configuration (Radinsky, 1973), or scavengers. The excessive wear on the posterior teeth of Epicyon supports the view that they literally devoured all nutritional parts of their prey. The large size of E. haydeni relative to the size of the Screw Bean ungulates, especially the abundant and diminutive antilocaprid, would have permitted the consumption of most of the skeleton. This would well explain the taphonomy of the Screw Bean Local Fauna.
Munthe (1998) states that Epicyon (E. saevus) was a cursorial predator adapted to open habitats that were expanding progressively northward from the Southwest into the Great Plains, during the late Miocene. Munthe (1989: 93) notes that E. “validus” (now E. haydeni) was a heavy animal (see Wang et al., 1999: 2) presumably incapable of long distance running, and implies that E. “validus” hunted as an ambush stalker. The metacarpal and metatarsal elements presented in Wang et al. (1999: 262–263) do not suggest a cursorial animal. The humeral fragment, TMM 42247-86, tentatively referred to E. haydeni (and the humerus illustrated by Munthe, 1989, F:AM 67603, as E. “validus,” [= E. haydeni], Jack Swayze Quarry, Clark County, Kansas, early Hemphillian [see Wang et al., 1999: 256]) lacks the narrowed and “rigid” distal articulation, a cursorial specialization, thus is somewhat bear like. Munthe failed to compare E. “validus” with bears, which is curious, because Munthe notes that E. “validus” died out when the bear Agriotherium appeared in North America.
Vulpes Frisch, 1775 Vulpes sp. Figure 9.7H–O; table 9.4
Leptocyon, cf. L. vafer (Leidy): Stevens and Stevens, 1989: 79.
Referred Material:
TMM 42247-250, partial rostrum with nasals and P1–3, and protocone of P4, associated with lower jaws with canine–p4, and TMM 42247-183, lower jaw fragment with premolar alveoli and a partial m1.
Stratigraphic Occurrence:
Bench 2, Liles measured section (fig. 9.2F), Banta Shut-in formation, Big Bend National Park, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description:
The partial rostrum preserves some of the nasals and pieces of the maxillary bones with the premolars (fig. 9.7L, M). The well-spaced premolars lack distinct accessory cusps. TMM 42247-183 preserves the alveoli for the canine, the roots for p2–4, and a damaged m1. The premolar spacing between TMM 42247-250 and TMM 42247-183 is very similar. The Screw Bean lower jaws are shallower and have more attenuated and more widely spaced premolars (fig. 9.7H–J, N, O) than in Leptocyon vafer examined, including the type. The Screw Bean specimens closely match homologues in Vulpes macrotis, the desert kit fox (see Hall and Kelson, 1959; Thornton and Creel, 1975) that lives in Trans-Pecos Texas today.
The m1 trigonid crown is broken away for TMM 42247-183, but its talonid is adequately preserved (fig. 9.7H–J). The metaconid is an enlarged, divergent cusp. The low lingual valley between the entoconid and metaconid supports a very small cuspule. The anteroposteriorly elongated talonid basin is rimmed lingually by a low entoconid and labially by a tall hypoconid. A slight cingular heel occurs between the entoconid and hypoconid.
Discussion:
The Screw Bean fox was previously identified as cf. Leptocyon vafer. L. vafer continued into the earlier Hemphillian (Whistler and Burbank, 1992), or to the onset of the late Hemphillian (Munthe, 1998: 139). The type, USNM 126 (see Leidy, 1869: 29, pl. 1, fig. 11), is a pair of lower jaws with teeth, from the late Barstovian “Niobrara sands,” Niobrara River or Burge Fauna (Webb, 1969: 35), near what is now Valentine, Nebraska. Matthew (1918) distinguished Leptocyon from Vulpes on the basis of the heel of m1 with a low entoconid crest that is indistinctly divided into two cusps, m2 with a low, shelflike paraconid, and very small m3. Gregory (1942: 348) did not believe that the features set forth by Matthew to separate Leptocyon from Vulpes had generic significance, but Webb (1969: 40) disagreed. Gregory observed (verified by MSS) that low entoconids are often present in modern red and kit foxes as an atavistic trait. Tedford (written commun., Mar. 20, 1989) noted that within Leptocyon-Vulpes, the m1 entoconid progressively becomes a more distinct cusp in younger species.
The lower jaw of the Screw Bean fox is relatively longer and more slender than in the type of Leptocyon vafer, and a referred Lapara Creek specimen (Wilson, 1960), TMM 30896-241, used for comparisons. Additionally, the Screw Bean premolars are more separated, and the jaw is shallower. The reduced accessory cusps and the more slender jaw are features derived in the direction of Vulpes; thus the Screw Bean fox, despite similar small size, is not referable to L. vafer. Our vulpine is interpreted to be a small species of Vulpes. Tedford et al. (1987: 190) note that undoubted Vulpes makes its first appearance in the Hemphillian as one of the taxa that defines the Hemphillian North American Land Mammal Age. Munthe (1998: 138), however, extends Vulpes back to the early Clarendonian; Leptocyon and Vulpes have often been used interchangeably for Clarendonian and Hemphillian vulpines.
Radinsky (1973: 182) notes on the basis of brain endocasts that Leptocyon is the ancestor for all living canids, not Tomarctus as previously believed (also see Webb, 1969; Martin, 1989; Tedford and Wang, 1995). One species is best recognized, V. stenognathus Savage (1941: 694), based on lower jaws from the Optima Fauna, late Hemphillian. Munthe (1998: 133) recognized “V. shermanensis” Hibbard, from Edson Quarry, Sherman County, Kansas. “V. shermanensis” also has been reported for the Coffee Ranch Fauna (Dalquest, 1969, 1983) but Schultz (1977) identified the Coffee Ranch fox as V. stenognathus, implying synonymy.
Teeth of Vulpes stenognathus are very similar in size and structure to the red fox V. vulpes, as noted by Savage. Most notable to Savage was the seemingly great depth and compression of the jaw of V. stenognathus, which influenced the choice of the name, but the narrowness and jaw depth is probably an artifact of crushing. Savage also compared the Optima fox with? Vulpes sp. (Stirton, 1936: 170; MacFadden et al., 1979), from the Coffee Ranch–Hemphill Fauna, Texas, and noted that the Coffee Ranch Vulpes is larger, but has a shallower and less compressed jaw, in keeping with a less deformed ramus. Both V. stenognathus and “V. shermanensis” are substantially larger and stratigraphically younger than the Screw Bean fox, and it is unlikely that these are the same species. The small size, and the relatively low entoconid in our specimen suggests that the Screw Bean fox is a primitive Vulpes within the Leptocyon-Vulpes transition. To date, no such earliest Hemphillian species have been described.
Eucyon Tedford and Qiu, 1996 cf. Eucyon sp. Figure 9.7C, D
Unidentified canid, large: Stevens and Stevens, 1989: 79.
Referred Material:
TMM 42247-46, left jaw fragment with alveolus for p3, a partial p4, and anterior alveolus for m1.
Stratigraphic Occurrence:
Found as float derived from the top of Bench 2, Liles measured section, Bench Member, Banta Shut-in formation (fig. 9.2F), Big Bend National Park, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description:
The jaw fragment (fig. 9.7C, D) matches the size (labial jaw depth at p4, 20.02) and has the alveolar spacing (p3–4, 24.3**) of large individuals of the modern coyote, Canis latrans Say. The p3 alveoli are well separated and indicate an anteroposteriorly attenuated and relatively narrow tooth. The partially preserved p4 (13.8* × 6.5*) is larger than p3, and is constricted slightly medially as in the coyote. m1 is not preserved, but part of its anterior alveolus indicates an enlarged, anteroposteriorly aligned, typically canid m1.
Discussion:
TMM 42247-46 represents a Canis-like jaw fragment. The specimen is much larger than homologues in the Screw Bean Vulpes sp. TMM 42247-46 is not referable to the borophagines Epicyon haydeni, also present in the Screw Bean assemblage, or Carpocyon Webb because the p3 and p4 alveoli are well spaced and do not indicate foreshortened teeth.
The Canini began their adaptive radiation in the late Clarendonian–early Hemphillian (Tedford and Qiu, 1996: 36; Munthe, 1998: 140) from within Leptocyon of North America. True Canis (and closely related genera) is not defined on the dentition, conservative, but on inflation of the frontal sinuses and on minor features of the occiput (Tedford and Qiu, 1996). The amount of sinus inflation, however, can vary from very little to considerable, at least among C. latrans, the coyote. Canis may have evolved in Eurasia from canid immigrants originally from North America, at about 3 Ma; the derived descendants returned to North America as a part of the global dispersal event of modern dogs (Flynn et al., 1991: 260). C. lepophagus (Johnston, 1938) and Canis ferox (Miller and Carranza-Castañeda, 1998) witness the Blancan arrival of these archaic, coyote-like Canis in North America.
Canis lepophagus, late Blancan Cita Canyon beds, Randall County, Texas, is slightly smaller than the modern coyote. Johnston (1938) provides no dental measurements. Enlargements of Johnston's plate 3, figure 1, to natural size, indicate that the Screw Bean jaw fragment is similar in size to WTM 881. The age difference between the canine from the Cita Canyon and Screw Bean faunas, and the timing of the Blancan origin of Canis, however, imply that these species are not the same.
“Canis” ferox is based on a skull and lower jaw, IGM 1130, of a coyote-like dog from near Rancho Viejo, State of Guanajuato, Mexico, late Hemphillian. The Screw Bean jaw fragment is larger than the jaw figured by Miller and Carranza-Castañeda (1998). The size difference and the presumed differences in age between the Screw Bean and Guanajuato specimens detract from identifying our jaw fragment as “C.” ferox.
Older Hemphillian “coyote”-like canids traditionally have been called Canis davisi Merriam (1911). The type of “C.” davisi, M1–2, UCMP 545, came from the late early Hemphillian Rattlesnake beds, John Day Basin. Merriam (1911) describes the red fox-sized upper teeth as generally similar to, but smaller than in the modern coyote. Shotwell (1970) reports “Canis” davisi from the Little Valley and Juniper Creek local faunas, Hemphillian, Juntura Basin, Oregon. Shotwell's (1970) specimen, UO 26742, is smaller than specimens of C. latrans and our Screw Bean jaw fragment. “C.” condoni Shotwell (1956) is based on a lower jaw that retains m2 from the McKay Reservoir fauna, Oregon, late Hemphillian. Shotwell (1970) later thought that “C.” condoni may represent “C.” davisi, and this is accepted by Tedford and Qiu (1996; but also see Munthe, 1998). “C. condoni” is about the size of Vulpes vulpes, the modern red fox (Shotwell, 1956), thus “C. condoni” is considerably smaller than TMM 42247-46.
Tedford and Qiu (1996) designated “Canis” davisi the type species of Eucyon, and clarified the distinction between E. davisi and true Canis (mainly lack of inflation of the frontal sinuses, and minor details of the occiput). The Screw Bean canid and E. davisi are approximate contemporaries. E. davisi, however, is smaller than the Screw Bean dog (see Tedford and Qiu, 1996), thus probably not the same species. The Screw Bean dog is near the size of the coyote-like, considerably younger E. zhoui (ca. 4 Ma) of China. It is apparent that a moderate and a larger species of nonborophagine dog coexisted in the early early Hemphillian in North America. Due to incompleteness, the Screw Bean canine is tentatively placed in Eucyon.
INFRAORDER ARCTOIDEA FLOWER, 1869
PARVORDER MUSTELIDA TEDFORD, 1976
FAMILY MUSTELIDAE FISCHER DE WALDHEIM, 1817
SUBFAMILY MEPHITINAE BONAPARTE, 1845
Buisnictis Hibbard, 1950 Buisnictis chisoensis, new species Figure 9.8A–K; table 9.5
Buisnictis, n. sp., extinct spotted skunk: Stevens and Stevens, 1989: 79.
Hypodigm:
Type, and TMM 42247-27, a partial skull with left I3–M1; TMM 42247-134, the posterior part of a skull; and TMM 42247-200, a partial lower jaw with partial m1–m2.
Diagnosis:
Buisnictis chisoensis is distinguished from B. burrowsi Skinner and Hibbard, late Blancan, by much more distinct m1 metaconid, a more basined talonid, and much larger size; from B. meadensis Hibbard, and B. breviramus (Hibbard), early Blancan, by a smaller P4 paracone, slightly wider angle between the lingual sides of P4 and M1, M1 with slightly less expanded talon, a relatively wider m1, an anteroposteriorly shorter m2, and a slightly larger size; differs from Martinogale alveodens Hall (1930:147),? late Clarendonian–late Hemphillian, by teeth with “smoother” enamel contours, lack of cingula, jaw with wider ascending ramus, smaller and more anteriorly situated angular process, lack of a “bulbous” symphysis, and larger size; and differs from Spilogale rexroadi, early Blancan, by much larger size, a much smaller P4 paracone, and by m1 without a deep notch between the metaconid and entoconid.
Etymology:
chisoensis; chisos, for the Chisos Mountains, Big Bend National Park, Texas, and ensis, Latin for “from” or “of.”
Stratigraphic Occurrence:
From just below and from the top of Bench 2, Liles measured section (fig. 9.2F), Bench member, Banta Shut-in formation, Big Bend National Park, Brewster County, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description and Comparisons:
The skull of Buisnictis chisoensis is rather low (fig. 9.8A–C), with a slightly depressed rostrum, a short and broad face, nasal tips terminate in line with the canines, and the external nares are small. The internal nares end slightly ahead of the posterior margin of M1. The rostral sutures fuse in early ontogeny. The postorbital processes are very small, postorbital constriction is rather narrow, and the type has a slight sagittal crest. Two other somewhat smaller partial skulls, TMM 42247-27 and TMM 42247-134, lack sagittal crests. The mephitine auditory bulla is small but well rounded. The mastoid sinus invades the mastoid bone laterally (fig. 9.8A, B), as in extant Spilogale pygmaea Thomas (see Van Gelder, 1959: fig. 47).
It has not been practical to separate the left jaw from the type, because the teeth are in tight occlusion, but the right jaw has been freed. The dental formula is I3/3, C1/1, P2–4/2–4, M1/1–2. The relatively small upper and lower incisors form a nearly transverse row (fig. 9.8B). I3 is the largest incisor. The ovoid upper canine has a small posterior tubercle, and a very slight internal cingular swelling. Cingula are absent, as is typical among the living skunks Spilogale, Mephitis, and Conepatus. The double-rooted P2–3 increase in size posteriorly, and have small heels.
The P4 is sectorial and musteline-like (fig. 9.8F, K), with a small, anteromedially situated protocone in the two specimens in hand. Protocones, however, can vary from much enlarged (see MU 1110) to moderately small (MU 1405) within modern Spilogale putorius. The carnassial blade is long as in Spilogale, not foreshortened as in extant Mephitis and Conepatus, and the valley between the paracone and metacone is unconstricted, the parastyle is small, and the paracone is the tallest cusp. P4 lacks cingula. The angle between the posteromedial border of the P4 and the anterior margin of M1 in B. chisoensis is about 50°, about 45° in Blancan Buisnictis, and about 10–15° in living Spilogale. A maxillary from the Beck Ranch assemblage, Scurry County, Texas, early Blancan, identified as S. rexroadi by Dalquest (1972), MUC:FV 8349 (as recatalogued within the museum collections to conform with the published specimen number), has an angle of about 10°, hence is similar to modern Spilogale. MUC:FV 8349, however, has a relatively narrower P4 than in extant S. pygmaea, thus more closely resembles the Screw Bean taxon.
M1 is broader transversely than long anteroposteriorly, has an expanded talon, a moderately constricted “waist,” low cusps and crests, and has a wide labial shelf (fig. 9.8F, K). The M1 is proportionally similar to Blancan Buisnictis and modern Spilogale, but has a more expanded talon than the Beck Ranch specimen mentioned above.
The lower incisors are small and crowded, and the canine is a stout, upright tooth with a small posterior tubercle, and a very weak lingual cingular swelling. The crowded, doubled-rooted p2–4 increase in size posteriorly. P2–3, and presumably p4, have expanded heels and are slightly oblique to the dental axis, as known (fig. 9.8G, H). P2–3 lack cingula, and a lingual-medial cuspid is not developed.
The m1 has a well-developed metaconid that is set slightly posterior to and diverges from the higher protoconid, and an accessory rootlet beneath the protoconid, like that commonly seen in modern skunks (fig. 9.8H), and various other modern and fossil mustelids. The trigonid occupies about 65% of the crown, and the tooth is relatively as broad as in Spilogale. The m1 of B. chisoensis, however, is relatively broader than in Blancan Buisnictis. The m1 talonid is semi-basined and rimmed by a trenchant, dominant hypoconid, and displays a distinct but subordinate entoconid. The m2 is relatively large, oval, with a flattened crown and a distinct protoconid cusp.
The lower jaw of Buisnictis chisoensis, similar to jaws of living Spilogale pigmaea, has a slightly convex ventral border, a short, anteriorly placed angular process, and a broad ascending ramus (fig. 9.8D). Multiple mental foramina occur below p3.
Discussion:
Many of North American later Tertiary small mustelids are based on isolated teeth or lower dentitions. Usual distinctions between them depend on differences in gross size, the morphology of the phylogenetically conservative m1, and differences in age. The Screw Bean specimens provide an opportunity to associate upper with lower teeth, and to attribute teeth with complete cranial material.
Weasels have retained the sectorial P4, but developed a more hypercarnivorous m1 through reduction or elimination of the metaconid, accompanied by simultaneous development of a trenchant talonid through loss of the entoconid. Among skunks, the M1 is a broad, low, crushing tooth, whereas in weasels the M1 usually is small. There is no doubt that the Screw Bean mustelid is a skunk. Based on cranial and lower jaw morphology, the taxon could be placed within Spilogale, extant spotted skunks, but the sectorial P4 bars this placement.
Isolated upper jaws of fossil skunks sometimes have been regarded as weasels because of their sectorial P4. Hibbard (1941a: 340) described a partial lower jaw from Rexroad locality 3, Meade County, Kansas, with a mephitine m1 as Brachyprotoma breviramus. The species was recharacterized (Hibbard, 1941b) as having crowded, oblique premolars, on the basis of alveoli. He thought that B. breviramus represented the smallest species of the Pleistocene genus, Brachyprotoma Brown. Hibbard (1950: 164) based the genus Buisnictis on B. meadensis, known from a musteline-like P4, UMMP 25769, from the early Blancan Rexroad Formation at Fox Canyon, XI Ranch, Meade County, Kansas. Because he knew at that time that B. meadensis had a weasel-like P4, he did not compare the type with the mephitine lower tooth of Brachyprotoma breviramus. Hibbard (1950: 165) then reported as Brachyprotoma breviramus a second lower jaw with p4–m1, UMMP 25771, from Fox Canyon. Later recovery of the anterior part of a skull with a lower jaw, UMMP 30242, from the Fox Canyon quarry convinced Hibbard (1954a: 224–226) that Buisnictis meadensis and Brachyprotoma breviramus were the same species. The Rexroad skunk then became Buisnictis breviramus (Hibbard). Fieldwork in the lower part of the Rexroad Formation, Buis Ranch, Beaver County, Oklahoma, resulted in the recovery of Buisnictis schoffi Hibbard (1954b: 344), based on a partial maxilla, UMMP 30196. Hibbard compared B. schoffi with B. breviramus (which at that time included the sample from Fox Canyon) and noted minor differences.
Bjork (1970), prompted by the recovery of Buisnictis in the early Blancan Hagerman Local Fauna, concluded that the Fox Canyon specimens are more similar in size to the Buis Ranch collection than to B. breviramus from the slightly younger Locality 3 of the Rexroad Formation. Bjork removed the Fox Canyon specimens from B. breviramus and placed them in B. schoffi. The reader, however, is reminded that the Fox Canyon locality produced the type of B. meadensis. If Bjork (1970) is correct in equating the Fox Canyon and the Buis Ranch populations as an older, single species, then the composite population must be called B. meadensis, as recognized by Skinner and Hibbard (1972) and Baskin (1998a). This invalidates the species “schoffi.”
Skinner and Hibbard (1972: 110) then added Buisnictis burrowsi, from the early Pleistocene, late Blancan Keim Formation, Brown County, Nebraska. The valid species of Buisnictis thus became B. meadensis Hibbard, earliest Blancan, B. breviramus (Hibbard), late early Blancan, and B. burrowsi Hibbard and Skinner, late Blancan. The report of B. chisoensis, if correctly assigned, not only extends the range of Buisnictis downward, but also demonstrates the dental stereotypy of the genus and the antiquity of the Spilogale-like bulla. When Baskin (1998a) prepared his report, B. chisoensis from the Screw Bean Local Fauna had not been described, and Baskin questioned the Screw Bean “Clarendonian” occurrence of Buisnictis cited by Stevens and Stevens (1989: 79).
Hibbard (1954b: 348) stated “It is also probable that Martinogale, Spilogale, and Mephitis had a common ancestor in the Lower Pliocene” (at that time the Clarendonian was regarded “Lower Pliocene”). Baskin (1998a: 170) believes that the living North American skunks Spilogale, Mephitis, and Conepatus may descend from within a ca. 4.5 Ma immigration event of mephitines from Eurasia into North America in the early Blancan, initiated from Promephitis. Baskin suggested that Buisnictis, Plionictis, and Martinogale are derived from an earlier Clarendonian immigration event, derived from Mesomephitis. The occurrence of the Spilogale-like Buisnictis in the Screw Bean assemblage suggests that at least Spilogale of modern skunks is derived from within the Clarendonian immigration event.
The type of Martinogale alveodens Hall (1930) came from Edson Quarry, Sherman County, Kansas, late Hemphillian. Dunkle (1938) reported a more complete lower jaw, and upper teeth have been reported by Harrison (1978: 43). Hall (1930) thought that Martinogale alveodens was a musteline possibly ancestral to Mustela, the modern weasel. Dunkle concluded that the characteristics that separate Martinogale from Mustela are precisely the features that relate Martinogale with Spilogale. Harrison (1978: 42) notes that the accessory rootlets beneath the protoconid of m1 of M. alveodens provide unequivocal evidence to support a close relationship of M. alveodens with Spilogale, Mephitis, and Conepatus. Accessory rootlets, however, occur commonly in other mustelid subfamilies, thus they lack phylogenetic significance. Baskin (1998a: 160) states that Martinogale is related closely to Buisnictis.
Whistler and Burbank (1992) report Martinogale from the Epicyon aphobus/Hipparion forcei Zone, Dove Springs Formation, Ricardo Fauna, California, late Clarendonian. Baskin (1998a) identifies the fossils as M. alveodens, and mentions a skull with associated jaws. The specimen, LACM 56230, from locality 3776, consists of the ventral part of a basicranium and a fragmentary rostrum, associated with dentaries. The rostrum contains one P4 and both M1s. The dentaries contain p2–m1. LACM 56230 is a quarter to a third smaller than B. chisoensis, the M1 has a narrower talon, m1 has a distinct labial cingular rim, a taller metaconid, the lower premolars have sharper crests, broader and more cingulated heels, and p3–4 have a distinct medial-lingual cuspid, features that B. chisoensis lacks. The entotympanic does not appear to be as inflated as in B. chisoensis. The Ricardo species is a Martinogale that is surely in line with M. alveodens, but has less developed cingula and less sharp dental crests and ridges. B. chisoensis, however, is not placeable in Martinogale. The main obstacles barring a relationship is the prominent cingula, sharper dental crests and ridges, lower premolars with lingual-medial cuspids, different shaped mandible, and smaller size of Martinogale. The two genera were clearly separate as early as the late Clarendonian–earliest Hemphillian.
If correctly identified, Buisnictis chisoensis is the only Buisnictis for which the basicranium is known. There is no doubt that the skull is specialized toward Spilogale. Van Gelder (1959: 246) notes that among living skunks, only Spilogale exhibits lateral inflation of the mastoid sinus, so much so as to exceed in some individuals the bizygomatic diameter, almost achieved in the type of B. chisoensis. The oldest known skunk that has a Spilogale-like dentition is S. rexroadi Hibbard, first described from the Rexroad Formation, early Blancan, Meade County, Kansas. S. rexroadi has also been reported from the Beck Ranch (Dalquest, 1972) and Blanco faunas of Texas (Schultz, 1977: 126), early Blancan. No skull is known for this species. The only reported P4–M1, MUCFV 8349 discussed above, comes from the Beck Ranch site, and shows a typically Spilogale-like P4, with its enlarged and posterolingual protocone.
Van Gelder believed probably correctly that Mexico was the center of origin for Spilogale, because he viewed S. pygmaea as the most primitive living species. The Screw Bean Local Fauna is near or within this realm. Some of the presumed “primitive” traits, the gracile skull and lack of a sagittal crest, probably are due to dwarfism. The smaller size of the mastoid sinus relative to S. putorius, however, is similar to Buisnictis chisoensis, and is probably a true primitive character state (as are the unbroken white stripes of the pelt). S. pygmaea is similar in size to S. rexroadi, but is smaller than Buisnictis chisoensis (see Van Gelder, 1959: 381). B. chisoensis probably was a predator of small vertebrates and insects; Spilogale is predominantly insectivorous. If so, the lineage that progressed toward Spilogale became more insectivorous and diminished in size.
Bjork (1970: 30) observed a progressive size decrease for Buisnictis through the Blancan. This trend is consistent with the larger size of B. chisoensis relative to the next younger B. meadensis. In turn, this implies that the ancestor for B. chisoensis was larger still, and would seem to eliminate the Ricardo species as a possible ancestor, as would the morphological differences noted above.
Martes nambianus Cope (1874) from the Pojoaque Member, Tesuque Formation (see Galusha and Blick, 1971: 108), Clarendonian, is now placed in Martinogale (Harrison, 1983; Baskin, 1998a). Harrison (1983: 27), however, ventured the suggestion that the broadened heel for p4 of M. nambianus, which separates M. nambianus from M. alveodens, might place M. nambianus with equal ease in Pliogale. M. nambianus resembles our species by lacking cingula and has premolars with more similarly expanded heels, but is as small as M. alveodens, thus much smaller than Buisnictis chisoensis.
Pliogale, the most generalized member of the Pliogale, Martinogale, Buisnictis clade (Baskin, 1998a: 170), lacks cingula, and m1 has an external accessory rootlet. Pliogale furlongi (Merriam, 1911: 249) is a small species from the late early Hemphillian Thousand Creek Local Fauna, Humboldt County, Nevada. Its m1 is small with a small metaconid that is separated from the protoconid posteriorly by a more distinct and deeper groove. P. manka Green (1956), Wolf Creek Local Fauna, early Clarendonian, is suitably large enough to have a possible relationship with Buisnictis chisoensis. P. manka, however, has a taller m1 with a relatively longer trigonid, shorter talonid, and taller and more trenchant hypoconid, within a more hypercarnivorous dentition that is perhaps in keeping with its greater age relative to Buisnictis. The immediate ancestry of Buisnictis might extend back to an unknown species of Pliogale allied with P. manka that inhabited the Mexican realm during the Clarendonian. The ultimate ancestry of these animals must be traced back to Eurasia.
If the above scenario is correct, Spilogale is a remote descendant of the Clarendonian immigration event, mentioned above. This in turn suggests that the origin of at least Spilogale, among the living skunks, is more remote than usually believed. Radinsky (1973) suggested, through study of brain endocasts, that skunks and skunklike animals are only remotely related to typical mustelids. Ledje and Arnason (1996) and Dragoo and Honeycutt (1997) recommend that the Mephitinae should be elevated to the rank of family, on the basis of molecular studies. The antiquity of the Spilogale-like mastoid process of Buisnictis chisoensis supports the view that skunks have had a long and separate adaptive history.
Martes Pinel, 1792 Martes sp. Figure 9.7A, B; table 9.6
Martes sp., extinct marten: Stevens and Stevens, 1989: 79.
Referred Material:
TMM 42247-75, detached right and left palatal fragments of a presumed single individual, with alveoli for left P3, and both P4–M1s.
Stratigraphic Occurrence:
From the top of Bench 2, Liles measured section, Bench member, Banta Shut-in formation (fig. 9.2F), Big Bend National Park, Texas; Screw Bean Local Fauna, early early Hemphillian, late Miocene.
Description:
No contact exists between the right and left palatal pieces of what must have been a single individual. The left fragment preserves the two alveoli for P3, a complete P4, and an M1 with a damaged metacone. The right fragment preserves the P4 and the M1 with a damaged talon. The teeth are little worn (fig. 9.7A). P4 is a relatively broad, sectorial tooth. The anteromedially placed protocone is a distinct cusp separated from the tall paracone by a deep saddle. The distinct parastyle is connected to the tall paracone by a slight ridge. The metacone, a low cusp that is only slightly taller than the paracone, is connected to the paracone by a concave ridge which forms the carnassial blade. The carnassial “notch” is broadly open, as is characteristic of mustelids. A lingual cingulum is present. The P4 closely matches that for the modern pine marten, Martes americana in size and morphology.
The M1 has a low protocone connected to the small protoconule by a ridge. The protoconule, in turn, merges with the anterior cingulum, which communicates with the paracone by a low ridge. The low paracone is set close to the posteromedial metacone as in Martes. The paracone is about twice the size of the metacone and is slightly taller. A narrow stylar shelf is present labial to the paracone. No metaconule is present. The hypocone is slightly convex upward in lingual view and is expanded anteroposteriorly and posteriorly, which gives a distinct figure-eight shape to the tooth. The M1, when seen in posterior view, is strongly flexed. The flexure places the talon and the apices of the protocone, paracone, and parastylar shelf practically on the same plane. The M1 is remarkably similar to that of extant Martes americana.
Although lower teeth are unknown for our Martes sp., m1 probably had a reduced metaconid. The anterior expansion and great flexure of the talon of the M1 usurps the space necessary to accommodate that cusp.
Discussion:
The Screw Bean specimen is a mustelid the size of males of Martes americana. Many fossil specimens have been referred to Martes in the literature, but most are lower teeth or jaws that are not directly comparable to TMM 42247-75. Furthermore, the substantial interspecific size variation that exists among the modern mustelids must be considered.
Martes originated in the Old World, and has been known in Eurasia from the Miocene onward (Franzen and Storch, 1975: 245; Pilbeam et al., 1977; Qi, 1985; Baskin, 1998a). Martes (sensu lato) is known in North America since the early Barstovian. Multiple extinctions and reintroductions from Eurasia certainly are possible. Some of the North American presumed Martes species are now placed in Plionictis Matthew (1924) or Sthenictis Peterson (1910; see Baskin, 1998a: 162–163). Plionictis, late Hemingfordian to early Clarendonian, is hard to distinguish from Martes (Baskin, 1998a). The well-developed parastylar shelf, limited anterior and posterior expansion of the talon of M1 of Plionictis, and greater age serve to distinguish the Screw Bean species from Plionictis. Sthenictis, late Hemingfordian through ?early Hemphillian, differs from Martes by a broader stylar shelf and less anteriorly and posteriorly expanded M1 talon (Baskin, 1998a).
The great similarity of TMM 42247-75 with the American pine marten, Martes americana, demands comparisons with Martes. Baskin (1998a) recognizes three Tertiary North American species, Martes glareae, Martes stirtoni, and M. oregonensis. Martes glareae (Sinclair, 1915), ranges from the early Barstovian Lower Snake Creek through the early Clarendonian Minnechaduza faunas (see Skinner et al., 1977: 347; Tedford et al., 1987). The type lower jaw is not directly comparable with our specimen, but is a M. americana-sized species. Sinclair (1915: 80) notes, however, that the metaconid of M. glareae is larger than in M. americana. Because our species probably had a reduced metaconid similar to that of M. americana, the Screw Bean species probably is not M. glareae.
Martes stirtoni Wilson (1968) is known from the early Clarendonian WaKeeney Local Fauna, Trego County, Kansas. M. stirtoni differs from the Screw Bean Martes sp. by smaller size, heel of M1 less expanded anteroposteriorly and with a distinct lingual cingular ridge, and by the presence of a well-developed metaconule. The teeth of the type are relatively narrower, have sharper cusps and ridges, and are more sectorial than in the Screw Bean Martes, and the talon of M1 in lingual view is more distinctly convex upward.
Martes oregonensis (Shotwell, 1970: 79) is based on a partial right jaw with p2–m2, from the Chalk Butte Formation (undivided), Hemphillian, Little Valley, northern Malheur County, Oregon. The type and the Screw Bean species are not directly comparable. M. oregonensis is similar in size to the modern M. americana, and a referred P3–4 are similar (Shotwell, 1970). Shotwell, however, notes that M. oregonensis differs most notably from M. americana by having a much larger m1 metaconid. This indirectly suggests that the M1 of M. oregonensis had a less anteroposteriorly expanded heel than either modern Martes americana or the Screw Bean species. M. oregonensis also has better developed cingula and accessory cusps. Based on limited information it seems unlikely that M. oregonensis and the Screw Bean species represent the same taxon. Other Martes specimens of more or less similar age as the Screw Bean Local Fauna have not been described in detail (see Shotwell et al., 1963: 51; Shotwell, 1970: 5; Voorhies, 1990: 136, 142). The Screw Bean species might be related to any one of these.
CONCLUSIONS
The establishment of an age for the Screw Bean Local Fauna has been difficult due to the fragmentary nature of most of the specimens that can be identified at best only to genus. Study of additional and some better preserved specimens suggests that the fauna, originally thought to be late Clarendonian (Stevens and Stevens, 1989) is Hemphillian in age. None of the index, or first-appearance taxa, however, used to define the Hemphillian Age of Wood et al. (1941) are present. An age of early early Hemphillian, late Miocene, is given to the Screw Bean Local Fauna for the following reasons.
Wood et al. (1941) used the first appearance of ground sloths to mark the beginning of the Hemphillian Land Mammal Age. Marshall et al. (1979) viewed the sloth Megalonyx (Oshkosh Local Fauna, upper part of the Ash Hollow Formation, Garden County, Nebraska) as the oldest North American edentate, and placed this occurrence at >9.3 Ma, extrapolating from a 9.3-Ma age determination for an ash in nearby Greenwood Canyon. Breyer (1981) regarded the Oshkosh and Greenwood Canyon local faunas as correlative, and as just predating a mid-Hemphillian extinction event of ≈7 Ma. Tedford et al. (1987) and Voorhies (1990) believe these faunas to be about 6–7 million years old. Wagner (1981) and Lindsay et al. (1984) record an older North American edentate, Pliometanastes, from the Siphon Canal Local Fauna, Disaster Peak Member, Mehrten Formation. The Siphon Canal assemblage has been dated at 8.19 Ma (Wagner, 1981; Lindsay et al., 1984; Tedford et al., 1987). These edentate occurrences place the Clarendonian/Hemphillian boundary of Wood et al. (1941) at about 8.2 Ma. Tedford et al. (1987) and Whistler and Burbank (1992) have lowered the Clarendonian/Hemphillian boundary to about 8.9 or 9 Ma (also see Wang et al., 1999). This predates the oldest known North American edentates by some 0.7–0.8 million years, mentioned in the explanation for figure 10.1 of Woodburne (1987).
The Clarendonian Land Mammal Age, and thus the Clarendonian Chronofauna (Webb, 1969), is based on the Clarendon Fauna from Donley County, Texas. The Screw Bean Local Fauna lived at a time when various members of that chronofauna were becoming extinct, and when some modern lineages were just appearing. The gradual demise of the Clarendonian Chronofauna was complete by the time of the mid-Hemphillian extinction event (or “major faunal discontinuity”) mentioned above at about 7 Ma. Later in the Hemphillian, at about 6 Ma, holarctic immigrants arrived from Eurasia, during a sea level low-stand. Some of these immigrants constitute the Hemphillian index taxa of Wood et al. (1941) and the Mio-Pliocene Chronofauna (see Janis et al., 1998a: 88). This faunal discontinuity has prompted the division of the Hemphillian into an early and a late interval (see Wagner, 1981; Lindsay et al., 1984). Neotropical immigrants (such as edentates) began to arrive from South America and/or the Antilles, as the Clarendonian Chronofauna declined, in the “early” Hemphillian. Because of the extension downward of the Clarendonian/Hemphillian boundary to about 9 Ma, and due to the absence of edentates older than 8.2 Ma, the “early” Hemphillian has been subdivided into an “early early” and a “late early” interval (Tedford et al., 1987; Schultz, 1990; Baskin, 1998b; Janis et al., 1998a). This cumbersome terminology is used in this paper.
To date, no edentate remains occur in the Screw Bean Local Fauna, an assemblage that occupies a strategic position within the invasion route envisioned for South American taxa moving north. However, edentate remains, although rare, are geographically widely distributed in late early Hemphillian faunas (Leidy and Lucas, 1896; Sinclair, 1906; Stirton, 1939; Schultz and Stout, 1948; Hirschfeld and Webb, 1968; Webb and Tessman, 1968; Shotwell, 1970; Schultz, 1977; Marshall et al., 1979; Tedford et al., 1987; Voorhies, 1990). The absence of edentates in the Screw Bean fauna is troubling if the fauna is younger than earliest Hemphillian.
Another consideration is the lack of oreodonts, negative evidence, in the Screw Bean Local Fauna in spite of diligent search. The oreodont Ustatochoerus Schultz and Falkenbach, abundant in Trans-Pecos Texas in the earliest Miocene (Stevens, 1977; Stevens and Stevens, 1989), is typical of the Clarendonian Chronofauna. Tedford et al. (1987: 194) note that Ustatochoerus probably became extinct at the end of the Clarendonian. Breyer (1981: 1214) regards the presumed Hemphillian oreodont occurrences in Nebraska as suspect. A derived Ustatochoerus, U. californicus, however, did extend into Member 6 of the Dove Spring Formation, Kern County, California, to about 8 Ma or slightly later, earliest Hemphillian (Whistler and Burbank, 1992). The lack of Ustatochoerus in the Screw Bean Local Fauna, thought to be real, suggests that the fauna lived after Ustatochoerus had become extinct.
The Screw Bean Local Fauna contains “early early” Hemphillian taxa that are late survivors of the Clarendonian Chronofauna, but contains none of the “late early” Hemphillian Eurasian immigrants (see Webb, 1969, 1983, 1989; Tedford, 1970; Tedford et al., 1987; Lindsay et al., 1984; Voorhies, 1990). Among the Screw Bean carnivores is the borophagine dog, cf. Epicyon haydeni, and the cats cf. Pseudaelurus sp. and cf. Nimravides catocopis. Epicyon did not survive the “early” Hemphillian (see Lindsay et al., 1984; Leite, 1990; Munthe, 1998). According to Breyer (1981), Epicyon became extinct during the “mid-Hemphillian extinction event,” centered at 6–7 Ma. Lindsay et al. (1984) show that Epicyon did not survive beyond the beginning of the late early Hemphillian, at what now would be regarded Chron 4r (Berggren et al., 1995; Cande and Kent, 1995). Nimravides catocopis extends from the late Clarendonian to the end of the early Hemphillian (Lindsay et al., 1984; Tedford et al., 1987). Pseudaelurus has a stratigraphic range from the late Hemingfordian to the late or latest Hemphillian (Martin, 1998).
The Screw Bean Vulpes sp. may represent a stem member of its lineage. Buisnictis and Martes are typical genera of the Mio-Pliocene Chronofauna, late Hemphillian–early Blancan. The mephitine Buisnictis, however, was not descended from an immigrant from Asia that shaped the Mio-Pliocene Chronofauna as suggested by Janis et al. (1998a). Buisnictis probably derived from an older immigration event (see Baskin, 1998a) that flavored the Clarendonian Chronofauna. The absence of Buisnictis in earlier Hemphillian faunas on the High Plains may result from the fact that Buisnictis originated far to the south, toward or within the Mexican realm. Spilogale, a presumed Buisnictis descendant, probably originated there.
Martes-like mustelids long have been in North America. The occurrence of Martes sp. in the Screw Bean Local Fauna supports the view that dentally modern martens have been present in North America longer than is usually supposed, contrary to Anderson (1984). The boreal or montane habit of modern Martes is inconsistent with the paleoenvironment envisioned for the Screw Bean Local Fauna. Multiple extinctions of Martes in North America, and reintroductions of ecologically different relatives from Eurasia is probable.
The fauna occupies a strategic location within North America for north–south and east–west faunal interchange, yet the fauna has a low taxonomic diversity. The impoverishment is thought to be in part the result of depositional style(s) and climatic factors. The lack of an oreodont, thought to be real, is important. The lack of edentate remains is less helpful because of their rarity in early Hemingfordian faunas in general. Given the presence of late members of the Clarendonian Chronofauna and the absence of: (a) oreodonts (disappearance by ≈8.5 Ma), (b) edentates (appearance by ≈8.2 Ma), and (c) characteristic Mio-Pliocene Chronofauna index taxa, the age of the Screw Bean Local Fauna is thought to be early early Hemphillian sensu Tedford et al. (1987) and others. Other members of the fauna, yet to be reported, support this age assessment.
Acknowledgments
Sincere thanks are due the following people. The late Dr. W.W. Dalquest, Midwestern State University, Wichita Falls, Texas, provided loan of a large sample of Spilogale putorius, loan of Beck Ranch specimens of Buisnictis and Spilogale rexroadi, and loan of selected Coffee Ranch specimens of Osteoborus cyonoides. J. Howard Hutchison, UCMP, Berkeley, California, graciously made available casts of Pliogale furlongi and Martes campestris. Dr. Philip D. Gingerich and Mr. William J. Ryan, UMMP, Ann Arbor, made available for study Dr. Claude W. Hibbard's collection of Buisnictis, and supplied a cast of the type of Martes stirtoni. Dr. Daniel H. Russell, Department of Mammalogy, AMNH, New York, provided measurements of Spilogale pygmaea. Dr. Robert J. Emry, Smithsonian Institution, provided loan of a cast of the type of Leptocyon vafer. Dr. John A. Wilson, Vertebrate Paleontology Laboratory, The University of Texas at Austin, made available for study a specimen of Leptocyon vafer from the Lapara Creek Fauna, and the laboratory is thanked for making available a cast of Martinogale alveodens, LACM 56230, provided by LACM. Dr. Richard H. Tedford, AMNH, made available a cast of the type of Osteoborus direptor, provided loan of specimens of Epicyon from Port-of-Entry Pit, and provided measurements of other Epicyon specimens under his care. Logan Ivy, Collections Manager, DMNH, is thanked for a cast of the palate of the type of Osteoborus pugnator.
Members of the staff, Big Bend National Park, deserve special mention. Among them are former superintendents James W. Carrico, Robert L. Arnberger, and José A. Cisneros, and the present superintendent, Frank Deckert. Mr. K. Mike Fleming, former Director, Mr. Vidal C. Davila, Jr., current Director, and Tom C. and Betty L. Alex, of Science and Resource Management, Big Bend National Park, have been especially helpful. Without their facilitation and interest, this paper would not have been possible. Dr. Wann Langston, Jr., is thanked for pinpointing the location of the fossil locality discovered by Mr. Jim Liles, former Chief Ranger, Big Bend National Park, who found the first of the Screw Bean fossils. Mr. Billy Pat McKinney, former rancher in the Big Bend, is thanked especially for flying one of us (JBS) over the study area in his airplane, and he and his wife Sadie Jo are thanked for many other kindnesses. Drs. Wilson and Ernest L. Lundelius reviewed all or parts of the manuscript and gave helpful suggestions. Outside reviewers also are thanked.
REFERENCES
Fig. 9.1. Index map of Texas, Trans-Pecos Texas, and the Big Bend, progressively enlarged. Major Neogene bolsons are indicated: A, Davis Mountains; B, Chinati; C, Bofecillos; and D, Chisos Mountains volcanic centers
Fig. 9.2. Measured sections that show the localities from which fossil Carnivora were recovered, and informal members and key beds of the Banta Shut-in formation, Big Bend National Park, Texas. The top part of section D correlates with the bottom part for section E. Location and orientation of the traverses used in measuring sections is shown on figure 9.3A. Universal Transverse Mercator (UTM) coordinates of the section-tops are shown on the insert, upper left. Explanation of the lithologic symbols appears in the insert, lower left. Use of the term “hyperconcentrated flow” follows Coussot and Meunier (1996), Chien and Wan (1999), and Wan and Wang (1994). Vertical exaggeration is 28.7×
Fig. 9.3. Geology of the Estufa bolson. A, Generalized outcrop of the Estufa Canyon and Banta Shut-in formations, and the location of selected measured sections, insert A. Measured section A, Light layer; B, Hell hole; C, Calcites; D, Comanche ridge; E, Tortoise leg; F, Liles; G, Dry camp; H, Cannon ball; I, SAS; and J, Boxing Day. Sections D, E, and F apply to figure 9.2. B, Generalized geologic map of the Estufa Canyon and Banta Shut-in formations with emphasis on the Banta Shut-in formation. Major faults are shown. Arrows indicate mean paleoflow directions at specific localities. Stippled, Lower member; white, Middle member, dashed, Red clay member, diagonal, Bench member, Banta Shut-in formation. Cross sections A–A′ to H–H′ coincide with the panels for the fence diagram, figure 9.4. C, Joint sets at the top of Comanche ridge, Hill 2362, Roys Peak Quadrangle Map, 1971. The calcite-filled fractures suggest latest Neogene or Quaternary left-lateral motion
Fig. 9.4. Fence diagram of the Estufa bolson shown in perspective, with emphasis on the Banta Shut-in formation. The location of the panels, A–A′ to H–H′ as in figure 9.3B
Fig. 9.5. Cf. Pseudaelurus sp., Screw Bean Local Fauna. A–D, TMM 42247-182b, distal end of medial metacarpal or metatarsal: A, lateral; B, dorsal (anterior); C, oblique; and D, condylar view. E–G, TMM 42247-114, lateral metacarpal or metatarsal: E, lateral; F, dorsal; and G, condylar view. H–K, TMM 42247-182a, proximal end of left metacarpal III: H, proximal articulation; I, dorsal; J, external; and K, ventral (posterior) view. L, M, TMM 42247-127, proximal end of radius: L, capitulum; M, anterior view. N, O, TMM 42247-237, proximal end of ulna: N, anterior; O, external view. P, Q, TMM 42247-144, distal end of right humerus, slightly restored (entepicondylar bridge appressed against shaft as preserved): P, anterior; and Q, medial view. R–T, TMM 42247-136, distal end of left humerus: R, anterior; S, ventral; and T, posterior view of trochlea
Fig. 9.6. Cf. Nimravides catocopis, Screw Bean Local Fauna. A, B, TMM 42247-33, posterior one half of left m1, protoconid, and metaconid: A, occlusal; B, labial view; arrow indicates carnassial notch. C, D, TMM 42247-138, partial right upper canine, lacking enamel: C, labial; D, cross section view where broken. E–H, TMM 42247-171, right proximal phalanx, pollex: E, proximal articulation; F, dorsal (anterior); G, distal articulation; and H, medial view. I–M, TMM 42124-21, distal one half of left humerus, slightly restored, arrows indicate directions of deformation: I, distal end of trochlea; J, medial; K, anterior; L, external; and M, posterior view
Fig. 9.7. Mustelidae and Canidae from the Screw Bean Local Fauna. A, B, Martes sp., TMM 42247-75, alveolus for left P3, left P4–M1; M1 slightly restored from the right side. C, D, Eucyon sp. TMM 42247-46, left jaw fragment with alveolus for p3, fragmentary p4, and anterior margin of m1 alveolus. E–G, P, Q, Epicyon haydeni: E–G, TMM 42247-28, right M1–2; P, Q, TMM 42247-86, distal end, left humerus: P, medial; and Q, anterior view. H–O, Vulpes sp.: H–J, TMM 42247-183, partial right lower jaw with alveoli for canine and premolars, and a partial right m1; K–O, TMM 42247-250, dorsal and lateral views of rostrum and views of P1–3, and canine, p1–4. Scale bars as indicated
Fig. 9.8. Buisnictis chisoensis, new species, Screw Bean Local Fauna, Big Bend National Park, Texas. A–I, TMM 24427-29, skull and jaws, type: A, dorsal; B, ventral; C, lateral view of skull; D, lateral view of right lower jaw; E, labial; and F, occlusal view of right I3–M1; G, occlusal; and H, labial view of i2–m2; and I, lingual view of m1. J, K, TMM 42247-27: J, labial; and K, occlusal view, I3–M3. A–D and E–K have different scale bars
TABLE 9.1 Measurements of cf. Pseudaelurus sp. Specimens at Texas Memorial Museum (in millimeters)
TABLE 9.2 Measurements of cf. Nimravides catocopis Specimens at Texas Memorial Museum (in millimeters)
TABLE 9.3 Measurements of Epicyon haydeni Specimens at Texas Memorial Museum (in millimeters)
TABLE 9.4 Measurements of Vulpes sp. Specimens at Texas Memorial Museum (in millimeters)
TABLE 9.5 Measurements of Buisnictis chisoensis Specimens at Texas Memorial Museum (in millimeters)
TABLE 9.6 Measurements of Martes sp. Specimens at Texas Memorial Museum (in millimeters)
