General

All OTUs discussed here are compared by means of the cranial and mandibular parameters (figs. 3–Figure 4Figure 56) advocated by participants of the 1981 Hipparion Conference held at the American Museum of Natural History, New York (Eisenmann et al., 1988). Features of the cranial morphology also are discussed. Characters of dental morphology focus on the height of crown at which they are developed, guided in part by the methodology explained in Woodburne (2003), which also allocated lower dentitions to crania of Merychippus insignis, otherwise not previously published.

Figure 4

Diagram showing location of cranial measurements and cheek tooth measurement methodology. A, dorsal view. B, posterior view. A and B after Eisenmann et al. (1988). C, measurement of length and width of upper cheek teeth and length and width of protocone. D, measurement of MSTHT from base to tip of mesostyle at labial surface of cheek tooth.

In addition to pertinent literature, evaluations of the morphology of the OTUs discussed here are based on specimens housed in the F:AM collections of the AMNH, with the hypodigm and geologic occurrence being specified for each.

Character States

Those used herein are comparable to, but in some cases different from, those presented by Bernor et al. (1988) and Hulbert and MacFadden (1991); they are also fewer in number. In those previous studies, as well as in this one and in Woodburne (2003), the morphology displayed by Parahippus leonensis is considered to be the plesiomorphic condition. Taxa utilized for this analysis, relevant localities (fig. 1), and stratigraphic settings (fig. 2) are indicated below and in table 2.

Cranial and Dental Parameters

See tables 1 and 2.

Nebraska

Cormohipparion merriami, n.sp., is represented in the Burge Fauna, of Nebraska, in the Burge Member of the Valentine Formation. Stratigraphically, this is the highest member in that formation. The lower three members (Cornell Dam, Crookston Bridge, and Devil's Gulch) contain specimens referred to C. quinni, with the Hurlbut Ash near the top of the Cornell Dam Member having been dated at about 11.6 ± 1.1 Ma (Skinner and Johnson, 1984: 252; revised from 13.6 ± 1.3 Ma in Boellstorff and Skinner, 1977). Woodburne (1996b) argued in favor of the older of these fission-track ages being more accurate. Perkins and Nash (2002) indicated an interpolated age for the Hurlbut Ash as 13.6 ± 0.1 Ma, which is consistent with the proposals of Woodburne (1996b). Swisher (1992) confirmed this age as 13.6 ± 0.21 Ma by ⁴⁰K-⁴⁰Ar dating of glass shards.

According to Skinner and Johnson (1984: 285), the Burge Member unconformably overlies the Devil's Gulch Member (fig. 7) and is composed of cross-bedded friable quartz sand with greenish-gray to olive-green clasts of clay. Tedford et al. (1987) considered the Burge Fauna to be of late Barstovian age, and this was followed by Woodburne (1996b). Whistler and Burbank (1992) correlated the Burge Fauna with that of the Cupidinimus avawatzensis/Paracosoryx furlongi Assemblage Zone of the Dove Spring Formation, California, the base of which is correlated to Chron C5A or about 12.5 Ma in the Berggren et al. (1995) time scale (fig. 2). Tedford et al. (2004) assigned the faunas formerly allocated to the latest Barstovian interval (Ba3; Tedford et al., 1987) to early Clarendonian (Cl1), as followed here and as correlated by Whistler and Burbank (1992).

The Valentine Formation is unconformably overlaid by the Ash Hollow Formation (Skinner and Johnson, 1984: 298), which contains faunas that range in age from early Clarendonian to late Hemphillian (Tedford et al., 1987: 169–170, fig. 6.2). The Ash Hollow Formation reaches a thickness of at least 80 m (Skinner and Johnson, 1984: 295) and is composed of a succession of massive, friable sandstone and interbedded arenaceous siltstone. Species of Cormohipparion occur in the lower two members of the Ash Hollow Formation, the Cap Rock Member and overlying Merritt Dam Member.

The Cap Rock Member contains the Minnechaduza Fauna of early Clarendonian age, with a species formerly assigned to Cormohipparion occidentale (described as Neohipparion occidentale by Webb [1969]). This material has been examined for the present report and determined to be comparable to Cormohipparion fricki, n.sp., from Texas (figs. 2, 7). The Cap Rock Member contains the Swallow Ash, near its base. Swisher (1992) reported a ⁴⁰Ar/³⁹Ar total fusion age of 12.18 ± 0.1 Ma for the Swallow Ash, herein considered more accurate than the fission-track age of 10.6 ± 0.2 Ma provided by Skinner and Johnson (1984: 296). Whistler and Burbank (1992: fig. 8) indicate that the Minnechaduza Fauna of Nebraska correlates with the lower part of the Epicyon haydeni/Hipparion forcei Assemblage Zone in California or within Chron C5N in the Berggren et al. (1995) time scale. This suggests that the Minnechaduza Fauna is contained within an interval of about 10.0–11.0 m.y., but this is surely too young. Based on the underlying Swallow Ash dated at 12.18 Ma, and the 11.5 Ma Davis Ash (see below) that overlies the Cap Rock Member, the Minnechaduza Fauna is estimated to be about 12 m.y. old.

The Merritt Dam Member of the Ash Hollow Formation contains the XMas-Kat, Machaerodus, and Hans Johnson quarries. Faunas from these quarries are of medial Clarendonian (Cl2) age (Tedford et al., 2004: fig. 6.2). The Davis Ash occurs stratigraphically below these fossil quarries, in the lower part of the Merritt Dam Member. The Davis Ash is thus both superpositionally above the Minnechaduza Fauna sites in the Cap Rock Member of the Ash Hollow Formation and below those of the XMas-Kat, etc., quarries. Swisher (1992) indicated a ⁴⁰Ar/³⁹Ar total fusion age of 11.5 ± 0.1 Ma for the Davis Ash. This age is followed here and replaces fission-track ages of 9.7 ± 1.2 Ma (on zircons) and 7.5 ± 2.2 Ma (glass shards) obtained by Izett (1975) and another glass shard fission-track age of 10.2 ± 0.7 Ma cited by Skinner and Johnson (1984: 297) from the same split sample of Davis Ash. Another ash, the Machaerodus Ash, occurs stratigraphically above the quarries of the extended XMas-Kat and Machaerodus channels. Skinner and Johnson (1984: 307) report a fission-track (glass shard) age of 9.95 ± 0.8 Ma for the Machaerodus Ash. Based on the above comparisons between ⁴⁰Ar/³⁹Ar total fusion ages and the fission-track results, it is likely that the Machaerodus Ash is about 10 Ma old.

The Hans Johnson Quarry is also considered equivalent to the XMas-Kat and Machaerodus quarries, even though the Machaerodus Ash is not recorded at the Hans Johnson Quarry (Skinner and Johnson, 1984: 297, 306–316, figs. 3, 16–17). On radioisotopic grounds, the XMas-Kat and Machaerodus (and Hans Johnson by correlation) sites are about 10 Ma old. Whistler and Burbank (1992: fig. 8) suggested that the XMas-Kat quarry faunas are about equivalent to the upper part of the Epicyon haydeni/Hipparion forcei Assemblage Zone in California, which is correlated to an interval in about the upper half of Chron C5N to Chron C4Ar in the Berggren et al. (1995) time scale. This suggests that the XMas-Kat and correlative sites in Nebraska are contained within an interval of about 11–9 m.y. In that faunas of early Hemphillian age appear to be as old as about 9 Ma (Woodburne and Swisher, 1995; Tedford et al., 2004), an age of about 10 Ma for the XMas-Kat and correlative sites seems plausible.

Cormohipparion occidentale (as here restricted) occurs in the medial Clarendonian faunas from the Xmas-Kat, Machaerodus, and Hans Johnson quarries in Nebraska. Cormohipparion matthewi, n.sp., also is represented in these sites.

South Dakota

Skinner and Johnson (1984) discussed the Big Spring Canyon and Hollow Horn Bear Quarry local faunas and indicate the Big Springs Canyon Local Fauna as being part of the Minnechaduza Fauna (see also Webb, 1969: 19). Hollow Horn Bear Quarry occurs on the north side of the Little White River drainage, in Todd County, South Dakota (Skinner and Johnson, 1984: fig. 2A, loc. 96), in the lower part of the undifferentiated Ash Hollow Formation (Skinner and Johnson, 1984: fig. 3, table 1). Locality information indicates that a thick white ash occurs stratigraphically above the Hollow Horn Bear Quarry, which is most likely the areally extensive Davis Ash (R. H. Tedford, personal commun., 2002), dated at 11.5 ± 0.1 Ma (Swisher, 1992).

If the Hollow Horn Bear Quarry is overlaid by the Davis Ash (figs. 2, 7), then the fossils from that site are somewhat older than those of the XMas-Kat quarries in Nebraska, in contrast to the correlation shown in Skinner and Johnson (1984: fig. 3). In that report, the Big Springs Canyon, the Hollow Horn Bear Quarry, and sites from the Thin Elk Formation of Harksen and Macdonald (1961) were all aligned with the present position of the Ed Ross Ranch Quarry (figs. 2, 7). The material from Hollow Horn Bear Quarry (F:AM 71880; fig. 8A) has a plesiomorphously much shorter lacrimal relative to the length of the POB (ca. 50%; table 3A), with the anterior tip of the lacrimal terminating about 5 mm posterior to the maxillo-jugal suture instead of reaching it (as seen in figs. 10A–B); it shows a somewhat more ovate protocone than in the type of C. occidentale (fig. 9D) and likely is lower crowned (unworn M1 MSTHT likely 55 mm; table 4B) versus a likely unworn MSTHT of 50–65 mm in the type (table 4A). This dental and cranial morphology is more compatible with that found in Cormohipparion fricki, n.sp. (see below).

Table 3

Lacrimal Length versus POB Length, Unworn Crown Heights, and Protocone Dimensions, Species of Cormohipparion A. Lacrimal length versus POB length

Specimens possibly from the Thin Elk Formation, F:AM 71872 (fig. 11) and AMNH 10869, show a more derived dental morphology (higher crowned [ca. 60 mm unworn MSTHT; table 4C), and have more elongate protocones (table 4C) than F:AM 71880 (table 4B) and are considered more plausibly allied with C. occidentale. If the phyletic correlation also is a temporal one, then it suggests that these materials are likely to be younger than those from the Hollow Horn Bear Quarry and elicits the proposal that the type material of C. occidentale is of about this age rather than as old as taxa from the Minnechaduza-equivalent interval (see Webb, 1969). The presence of abundant garnet crystals in both the Thin Elk Formation and from the Merritt Dam Member of the Ash Hollow Formation (there associated with the XMas-Kat and Machaerodus quarries) led Skinner and Johnson (1984) to propose a correlation based thereon, compatible with the stage of evolution of the C. occidentale specimens.

Figure 11

Cormohipparion cf. occidentale. F:AM 71872, ?Thin Elk Formation, South Dakota, early Clarendonian. A, lateral view of cranium showing tall-crowned upper cheek teeth. B, occlusal view of left upper cheek teeth.

Texas

As indicated in Tedford et al. (1987), the main F:AM collections are derived from the MacAdams and Gidley Horse Quarry sites in the Clarendon beds of the Texas Panhandle. Schultz (1977) provided the most recent summary of the geological setting of these sites. As shown in figure 2, the MacAdams Quarry is chronologically older than the Gidley Horse Quarry sample, although both are of medial Clarendonian age (Cl2; Tedford et al., 2004: fig. 6.2). According to the radioisotopic evidence cited above, the MacAdams Quarry is between 11.5 Ma and 12.0 Ma old, and the Gidley Horse Quarry is about 10 Ma old. These correlations are about 1 m.y. older in each case than suggested by Whistler and Burbank (1992: fig. 8). Cormohipparion fricki, n.sp., occurs in the MacAdams Quarry, and C. skinneri, n.sp., is represented in the Gidley Horse Quarry.

California

Whistler and Burbank (1992) described the biochronology and the radioisotopic and magnetostratigraphic geochronology of the Dove Spring Formation in the El Paso Mountains of Southern California (E, fig. 1). The Dove Spring Formation is about 1,800 m thick and contains the classical “Ricardo Fauna” of Merriam (1919). Whistler and Burbank (1992: fig. 4) indicated that C. occidentale, s.l., ranges through the Cupidinimus avawatzensis/Paracosoryx furlongi and Epicyon haydeni/Hipparion forcei assemblage zones, with C. occidentale s.l. correlated on the basis of the paleomagnetic data to an age of 12.6–8.8 m.y. in the Berggren et al. (1995) time scale. Whistler and Burbank (1992: fig. 8) indicate that the Cupidinimus avawatzensis/Paracosoryx furlongi Assemblage Zone correlates with the Burge Fauna of Nebraska, at its base, and about with (likely somewhat older than) the Minnechaduza Fauna, Nebraska, and the MacAdams Quarry fauna, Texas, at its top. In the present report, the Burge and Minnechaduza faunas are considered to be of early and medial Clarendonian age, respectively, following Tedford et al. (2004). In any case, the Cupidinimus avawatzensis/Paracosoryx furlongi Assemblage Zone is from about 11.8 to 11 m.y. old and, following Tedford et al. (2004), the Epicyon haydeni/Hipparion forcei Assemblage Zone ranges in age from about 11 Ma to 9 Ma. In the present report, the material from the Dove Spring Formation, which lacks crania, is referred to Cormohipparion sp. and has a range from about 11.8 m.y. to 10.4 m.y. (fig. 2).

Woodburne (2005) described an upper dentition referred to Cormohipparion sp. from the Devil's Punchbowl, Valyermo, California (V, fig. 1). This is considered to be of likely medial Clarendonian age.

Type Species

Hipparion occidentale Leidy 1856

Type Specimen

ANSP 11287, four left and one right upper cheek teeth (MacFadden, 1984: 162). In Osborn (1918: fig. 140) and Skinner and MacFadden (1977), RP2, RP3, ?RM1, RM2, and LP3 are represented.

Included Species

Species included in the subgenus Cormohipparion, revised from Hulbert (1988), are C. goorisi MacFadden and Skinner 1981 (early Barstovian, Texas); C. quinni Woodburne 1996b (late Barstovian, Nebraska and Colorado); C. matthewi, n.sp. (late medial Clarendonian, Nebraska); C. merriami, n.sp. (early Clarendonian, Nebraska); C. johnsoni, n.sp. (early Clarendonian, Nebraska); C. fricki, n.sp. (early medial Clarendonian, Texas and Nebraska); C. skinneri, n.sp. (late medial Clarendonian, Texas); and C. occidentale (late medial Clarendonian, Nebraska and South Dakota). Materials allocated here to Cormohipparion sp. occur in El Paso Basin faunas of medial Clarendonian age in California (fig. 2) and in the Devil's Punchbowl, Valyermo, California (fig. 1), also likely of medial Clarendonian age (Woodburne, 2005).

Species included in the subgenus Notiocradohipparion Hulbert 1988 are C. (N.) plicatile (MacFadden, 1984; Hulbert, 1988); medial Clarendonian through medial Hemphillian, Florida; C. (N.) ingenuum (MacFadden, 1984; Hulbert, 1988); medial Clarendonian through medial Hemphillian, Florida; medial Clarendonian, Texas; early or medial Hemphillian, Honduras; C. (N.) emsliei Hulbert, 1987; late Hemphillian to Blancan, Florida; late Hemphillian, Louisiana.

Geographic and Chronologic Distribution

Early Barstovian of Texas; late Barstovian of Colorado and Nebraska; early Clarendonian of Nebraska; medial Clarendonian of California, Nebraska, South Dakota, Texas, and Florida; Hemphillian to Blancan of Florida; late Hemphillian of Louisiana. As discussed below, Cormohipparion is not found yet beyond North America, contrary to Bernor et al. (2003).

Type Specimen

Hulbert (1988) gave a definition of the subgenus Notiocradohipparion but not of the sister taxon he nominated as the subgenus, Cormohipparion. Relative to Notiocradohipparion, the subgenus Cormohipparion is plesiomorphic in the retention of a short muzzle. The most primitive members of the subgenus Cormohipparion (C. goorisi and C. quinni) have the derived condition of a strongly pocketed posterior extent of the DPOF, and all members of this subgenus are derived in having the IOF located anteriorly, above the P3 or the P3–P4 boundary, and closely associated with the anterior or anteroventral edge of the DPOF. In this combination of characters, at least, members of the subgenus Cormohipparion differ from approximately contemporaneous other taxa: Parahippus leonensis, Merychippus insignis, M. shirleyi, “M.” primus, “M.” tertius, Parapliohippus carrizoensis, Acritohippus stylodontus, Scaphohippus sumani, “M.” coloradense, P. mirabilis (list from Hulbert and MacFadden, 1991: fig. 15; note that other species in this figure are known only from dentitions; Parapliohippus carrizoensis and Acritohippus stylodontus is from Kelly, 1995; Scaphohippus sumani is from Pagnac, 2006).

Type Locality

Little White River, South Dakota (Skinner and Taylor, 1967).

Age

Likely medial Clarendonian.

Definition of Cormohipparion

Note that this definition does not depend on the number of OTUs in C. occidentale, s.l. This definition is taken from Woodburne (1996b) and therefore is modified from MacFadden (1984) and Hulbert (1987): Mesodont to hypsodont hipparion. The mean TRL ranges from about 116 mm to 140 mm. The unworn or little-worn M1MSHT ranges from about 34 mm to 66 mm (table 3B). The DPOF is prominent with a relatively well-developed and usually continuous anterior rim, with the IOF located above P3, the P2–P3 boundary, or the P3–P4 boundary and consistently very close to the anteroventral tip of the DPOF. Posteriorly, this fossa has a well-developed and pocketed rim. In general, the fossa is oval or teardrop-shape in outline and situated far anterior of the orbit, resulting in a wide POB. The anterior tip of the lacrimal bone enters into the rear of the DPOF (primitive species) or does not reach the rear of the fossa (advanced species). The DPOF is lost or severely reduced in advanced species of C. (Notiocradohipparion), such as C. (N.). emsliei (Hulbert, 1987). The DP1 is relatively large and persistent into the adult condition, except in more derived species, where it tends to be smaller and not to persist into adult wear, or it tends to be limited to females. The protocone is isolated (except in P2) until very late wear, with a spur in plesiomorphic taxa. The protocone becomes ovate to elongate-oval in shape in evolved species. The P2 anterostyle is well developed; P2 is slightly (plesiomorphously) longer to much longer than other cheek teeth. The pli caballin is prominent; it is usually multiple (2 or more plis) in premolars and single in molars. Fossette borders are moderately to very complex (especially the opposing borders of pre- and postfossettes). The anterior border of the prefossette is increasingly complex in derived taxa. The lower cheek teeth generally possess protostylids (except C. goorisi).

Table 3 (Continued)

B. Unworn MSTHT heights^a (mm) for P3–M3 in Cormohipparion species

Cormohipparion occidentale (Leidy), 1856

figures 9–Figure 1011, 13; tables 2–Table 3Table 3 (Continued)Table 3 (Continued)Table 4{ label needed for table-wrap[@id='i0003-0090-306-1-1-t07'] }Table 4(Continued){ label needed for table-wrap[@id='i0003-0090-306-1-1-t09'] }Table 4 (Continued)Table 5Table 6Table 6 (Continued)Table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }8

Table 3 (Continued)

C. Protocone width/length ratio in Cormohipparion species

Table 5

Measurements (mm) of Crania of Cormohipparion occidentale, Late Medial Clarendonian, Nebraska and South Dakota

Table 6

Measurements (mm) of Diastema Length and Canine Position in Cormohipparion Species

Table 6 (Continued)

Type Specimen

ANSP 11287, four left and one right upper cheek teeth (fig. 9D). See Skinner and MacFadden (1977: 919, fig. 4D) and Osborn (1918: 176, fig. 140). According to Osborn (1918: 176), these are RP2–3, M2, and LP3. Another specimen, ?RM1 (Skinner and MacFadden, 1977), is not specifically discussed, nor is it part of the cast series (AMNH 10794) in the AMNH.

Type Locality

Probably from along Little White River, South Dakota, a possible correlate of the XMas-Kat quarries in Nebraska (Skinner and Johnson, 1984).

Age

Clarendonian, likely late medial Clarendonian.

Distribution

?Late medial Clarendonian of South Dakota; late medial Clarendonian of north-central Nebraska (XMas-Kat, Machaerodus, and Hans Johnson quarries).

Description of Type Specimen

This is based on casts of ANSP 11287 in the AMNH and on Osborn (1918: 177, fig. 140), Skinner and MacFadden (1977), and MacFadden (1984). The material is considered to represent four right cheek teeth and a single left upper molar, RP2, RP3, ?RM1, RM2 and LP3, with P2–M2 illustrated as reversed in figure 9D. As indicated in table 4A, the teeth are interpreted to be in about middle to one-third wear, with mesostyle heights ranging from about 30 mm (P2) to 41 mm (M2). The enamel pattern is of moderate complexity, with 3–5 plis on the anterior border of the prefossette, 4–7 plis on the posterior border of the prefossette, 3–5 plis on the anterior border of the postfossette, and 1 pli on the posterior border of the postfossette. The protocone is ovate to elongate-oval, nearly flat to slightly concave lingually, and nearly flat to slightly convex labially. The hypoconal groove is simple and open and lacks a hypoconal spur. The pli caballin is most complex in the premolars (2–3) and single in the only molar. The pli caballin is directed somewhat posterolingually in all of these teeth, except for M2, where it is slightly anterolingually directed. The pre- and postfossettes are united along their labial margins in P2. Based on the crown height of M2 (table 4A), the unworn MSTHT for these cheek teeth appears likely to have been in the range of 50–65 mm, based on other material referred to C. occidentale. The pre- and postfossette of P2 are confluent labially.

The type material was compared with specimens from F:AM locality Hollow Horn Bear Quarry, South Dakota, as potentially representative of the area from which the type materials are thought to have been collected. As discussed above, this site apparently is correlative with the Minnechaduza Fauna (figs. 2, 7), and, as described below, F:AM 71880 (fig. 8B) therefrom has a more plesiomorphic molar and lacrimal morphology compared with examples of C. occidentale. F:AM 71880 is herein referred to C. fricki, n.sp., as are the other specimens from the Hollow Horn Bear Quarry (table 4B).

Specimens possibly from the Thin Elk Formation, F:AM 71872 and AMNH 10869 (table 4C), are higher crowned than seen in C. fricki, n.sp.; show a more derived dental morphology (more elongate protocones) than F:AM 71880 (table 4B); and are considered to be chronologically younger than specimens from the Hollow Horn Bear Quarry. These same specimens are morphologically comparable to the type material of C. occidentale and suggest that both the Thin Elk Formation and the type material are younger than taxa from the Minnechaduza-equivalent interval (see Webb, 1969). If the Thin Elk Formation provenance of F:AM 71872 and AMNH 10869 can be verified, the presence of abundant garnet crystals in both the Thin Elk Formation and the Merritt Dam Member of the Ash Hollow Formation (there associated with the XMas-Kat and Machaerodus quarries) also would be consistent with the age interpretations explored here.

The type of C. occidentale also was compared with a sample of Neohipparion affine (following MacFadden, 1984), as this taxon also has an upper cheek tooth morphology in which an elongate, isolated protocone is prominent (contrary to species of Hipparion and Nannippus as summarized in MacFadden, 1984). Specimens with dentitions in crania that also preserve the facial morphology are represented, in part, by AMNH 141218 (fig. 12A–B) from the Hollow Horn Bear Quarry and by F:AM 111728 (MacFadden, 1984: fig. 66; fig. 12C–D), F:AM 111729 (MacFadden, 1984: fig. 67), and F:AM 132652–132656 from the MacAdams Quarry, Clarendon beds, Texas. As shown in figures 12B and 12D, the elongate and lingually concave protocone of these specimens is similar to that shown in the type of C. occidentale. These specimens all differ, however (table 4D), in the distinctly simpler fossette border morphology, the single pli caballin in premolars as well as molars, the anterior orientation of this pli caballin in premolars as well as molars, the more lingually slanted axis of the hypoconal groove in the premolars overall, and the generally less prominent pli protoconule of the premolars as well as the molars. The cheek teeth tend to be narrower at comparable stages of wear. Based on these specimens, as well as on other material referred by MacFadden (1984) to N. affine, the unworn MSTHT likely was about 55 mm (e.g., F:AM 108236 from the Eli Ash Pit, Cap Rock Member, Ash Hollow Formation, medial Clarendonian, of north-central Nebraska, with very early wear M3 being 52 mm tall, and an early wear M2 and P4 being 54 mm and 52 mm tall, respectively; MacFadden, 1984: fig. 68). F:AM 111728 (MacAdams Quarry, fig. 12E) shows that protostylids are present in p2–m3, as in C. occidentale (see below), even in relatively late wear (m1 metaconid and metastylid effectively confluent; MacFadden, 1984: fig. 69).

If the unworn MSTHT of the upper cheek tooth dentition of the type of C. occidentale was on the order of 60 mm, as seems likely, it would be beyond the range for all but the XMas-Kat materials and those from Gidley Horse Quarry, Texas, a comparison reinforced by the elongate protocones and fossette border complexity seen in the type material.

The comparisons above validate the operation taken by Skinner and MacFadden (1977) in referring a large sample of late Barstovian and Clarendonian equids to Cormohipparion and some of these to C. occidentale based on the associated dental features described above and the distinctive morphology of the DPOF. This also is consistent with the action taken by MacFadden (1984) in restricting the nomen Neohipparion to specimens referred by him to N. affine, N. coloradense, N. trampasense, N. leptode, N. eurystyle, and N. gidleyi.

Revised Diagnosis

Based on the type and referred material, within the subgenus Cormohipparion, C. occidentale is distinguished from C. goorisi and C. quinni (fig. 9) in larger cranial (and presumably body) size; in higher crowned cheek teeth; in having a more complex enamel pattern in the upper cheek teeth; in dP1 being reduced to absent; in lower cheek teeth with stronger protostylids; in a cranium with the DPOF having a generally teardrop shape (dorsoventrally much higher posteriorly than anteriorly); and in the lacrimal not reaching the rear of the DPOF.

With respect to other members of the C. occidentale group, C. occidentale, s.s., is distinguished in having 4–5 plis on the anterior border of the molar prefossette. The molar posterior border of the prefossette has 8–9 plis; the premolar posterior border of the postfossette has 2–3 plis; the molar posterior border of the postfossette has 2–3 plis; the molars usually have two plis caballin; and the orbit is posterior to M3. With a range of about 50–66 mm, the unworn MSTHT for P3–M3 is higher than that for all other species, possibly except C. skinneri, n.sp. (table 3B), a segregation strengthened by a comparison of unworn MSTHT for P4/M1 in these samples, as well.

In addition, and as compared in figures 13, 15, 18, 20, 22, and 24, C. occidentale has the most deeply incised nasal notch (no. 30) of all the other species, and the lacrimal does not extend anteriorly beyond the junction with the maxillo-jugal suture, in contrast to C. merriami, n.sp., and C. johnsoni, n.sp., but not C. fricki, n.sp., C. matthewi, n.sp., or C. skinneri, n.sp. As in C. matthewi, n.sp., the lacrimal in C. occidentale extends across about 60% of the POB (table 3A), greater than all other species except C. skinneri, n.sp., and C. johnsoni, n.sp. Cormohipparion occidentale also has a longer muzzle (parameter 1) than all members of the Cormohipparion group but C. merriami, n.sp. The 1 S.D. (standard deviation) range of the length of the DPOF (no. 33) in C. occidentale incorporates the (low) range of this feature seen in C. merriami, n.sp., and C. skinneri, n.sp., so these populations probably are not distinct in this regard. However, the DPOF is sharply shorter in C. matthewi, n.sp., and C. johnsoni, n.sp., and its 1 S.D. ranges are much smaller in C. fricki, n.sp. None of the other species shows as tall a DPOF (no. 35) as in the Machaerodus and Hans Johnson quarry samples of C. occidentale, and only C. johnsoni, n.sp., has as short a separation of the DPOF and facial crest (no. 36) as in those two quarry samples of C. occidentale.

Figure 15

Log-ratio diagram of cranial dimensions of Cormohipparion matthewi compared with C. occidentale, XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska.

Figure 18

Log-ratio diagram of cranial dimensions of Cormohipparion johnsoni compared with C. occidentale, XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska.

Figure 20

Log-ratio diagram of cranial dimensions of Cormohipparion merriami compared with C. occidentale, XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska.

Figure 22

Log-ratio diagram of cranial dimensions of Cormohipparion fricki compared with C. occidentale, XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska.

Figure 24

Log-ratio diagram of cranial dimensions of Cormohipparion skinneri compared with C. occidentale, XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska.

Referred Material

From Little White River, South Dakota. Medial Clarendonian.

From the Ed Ross Ranch Quarry, ?late medial Clarendonian, ?Thin Elk Formation, Bennett County, South Dakota. F:AM 71872, ♂ skull with right and left C1, P2–M3 (fig. 11).

From Loup Fork beds, ?Clarendonian. AMNH 10869, fragmentary skull with right and left dP2–4, right and left M1, RM2 erupting.

From Niobrara River, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska. Late medial Clarendonian.

XMas-Kat quarries. F:AM 71800, nearly complete cranium with right and left I1–3, C, P2–M3, adult female, M3 well worn. F:AM 71801, nearly complete cranium with right and left I2–3, C, P2–M3, sub-adult male, M3 just erupting. F:AM 71803, nearly complete cranium with L I2–3, right and left C, P2–M3, adult male, M3 erupted. F:AM 71806, partial cranium with right and left I1–3, P2–M3, adult female, M3 in medial wear. F:AM 71807, right and left palate with P2–M3, M3 in early wear. F:AM 71808, partial cranium with right and left P3–M3, M3 in early wear. F:AM 71809, partial cranium with right and left P2–M2, LM3, M3 well worn. F:AM 71810, partial cranium with right and left P2–M3, M3 well worn. F:AM 71813, nearly complete cranium with right and left I1–3, RC, right and left P2–M3, adult female, M3 well worn. F:AM 71814, badly crushed cranium with L I2–3, C, right and left P3–M3, adult male, M3 well worn. F:AM 71868, anterior cranium with R I2–3, right and left C, dP2–4, LM1–2 erupting, LP2–4 removed from crypt, juvenile male.

Machaerodus Quarry. F:AM 71831, snout and frontal regions preserved, with right and left I1, L12, right and left C, RdP1, right and left dP4, P2–M3, M1–2 erupting, lower mandibles with right and left i1, Ri3, right and left c, p2–m2, juvenile male, p2 and p3 erupting. F:AM 71832, nearly complete cranium with right and left I3, C, LdP1, right and left P2–M3, adult female, M3 well worn. F:AM 71834, snout and facial region with right and left I1–2, RI3, P2–M3, ancient female, M3 very well worn. F:AM 71835, palate with LP2, right and left P3–M3, adult, M3 in early wear. F:AM 71844, partial snout and palate with R C, P2, right and left P3–M3, adult male, M3 in early wear, and associated left and right mandibles with p2–m3.

Hans Johnson Quarry. F:AM 71855, laterally compressed cranium with right and left I1–2, LI3, C, RP2, right and left P3–M3, adult male, M3 well worn; mandible with right and left i1–3, c, Pp2–m3. F:AM 71856, nearly complete cranium with right and left I1–3, C, P2–M3, adult male, M3 worn. F:AM 71857, nearly complete cranium with right and left I2–3, dP1–M3, adult female, M3 well worn. F:AM 71858, nearly complete cranium with R I1, C, right and left dP1–M3, adult female, M3 well worn. F:AM 71860, partial cranium with L I3, right and left C, P2–M3, adult male, M3 very well worn. F:AM 71861, partial cranium with right and left P2–M3, adult female, M3 beginning wear. F:AM 71862, crushed cranium with right and left I1, right and left I2-3, C, dP1–M3, subadult female, M3 erupting.

Description of Referred Material

The cranium of C. occidentale is poorly represented in specimens from localities of medial Clarendonian age in South Dakota. The following description is based on material from the late medial Clarendonian XMas-Kat quarries, herein taken as the most representative of this species (see below). F:AM 71800 is a nearly complete and undistorted cranium of an adult female (fig. 10A). The skull is slightly crushed dorsoventrally but otherwise nearly complete. The parietal region is somewhat depressed; the maxillary overlaps by about 20 mm its normal contact with the nasal bone; and the lower suture of the lacrimal is obscured. The configuration of the face in this area is more accurately represented by F:AM 71801 (fig. 10B). Figure 10A illustrates the generally elongate configuration of the cranium (snout and facial region) typical of late Miocene hypsodont horses. The large DPOF (no. 33, table 5) characteristic of Cormohipparion species is clearly shown, as are its strongly demarcated borders. The anterior tip of the DPOF is closely associated with the infraorbital foramen, which is located above the anterior margin of P3 and at a level opposite the lower margin of the orbit, and exemplifies the anteroventral orientation of this fossa (state 0, character 5, table 1). Figure 10B demonstrates the extent to which the DPOF is pocketed in C. occidentale, reaching posteriorly about 10 mm to a point internal to the tip of the lacrimal bone; the anterior tip of the lacrimal is about 15 mm posterior to the rear of the DPOF. The lacrimal is not pointed anteriorly, but its anterior edge (maxillo-lacrimal suture) is nearly vertically oriented and approximately aligned with the maxillo-jugal suture. The nasal notch is well developed and extends posteriorly to a point above the anterior margin of P2. The anterior edge of the large orbit is located above the rear of M3 and is well separated from the DPOF by a wide POB (no. 32, table 5). As shown in table 5, the orbit tends to be somewhat longer (no. 28) than high (no. 29). The facial crest is well developed and extends anteriorly to a point above M1. Although crushed in F:AM 71800, F:AM 71801 shows that the frontal region is slightly domed anterior to the parietal crest. The occipital condyles are slightly below the level of the tooth row in lateral view, and the paroccipital processes are strongly developed.

Crania from the Machaerodus and Hans Johnson quarries are comparable to those from XMas-Kat. Figure 13 shows the log-ratio values for the C. occidentale samples from the Machaerodus and Hans Johnson quarries relative to that from XMas-Kat. Parameter 4 is a measure of the basicranium and thus of fundamental cranial size. The correspondence between the samples indicates that they are all about the same size. Although the Machaerodus Quarry occasionally is represented by only one specimen (in which case it is assumed that the dimension represents the log mean for its sample), the relatively close correspondence of the log means of all three samples indicates that range of variation (± 1 standard deviation) would overlap extensively. Parameters in which the S.D. range of the XMas-Kat sample overlaps the other two are no. 22 (inion height), no. 30 (nasal notch), no. 34 (IOF relative to rear of DPOF), and no. 37 (IOF height relative to alveolar border). Parameters in which the S.D. range of the XMas-Kat sample is overlapped by the other two (assuming the S.D. range for the sample represented by a single Machaerodus specimen is comparable to that from the Hans Johnson Quarry) are no. 1 (muzzle length), no. 3 (postpalatal length), no. 12 (choanal width), no. 13 (palatal width), no. 19 (transglenoid width), no. 25 (muzzle height; assumes a greater range for the Hans Johnson sample), no. 31 (facial length; log mean only in Machaerodus Quarry; includes no. 2 in part), no. 32 (POB length), and no. 39 (lacrimal tip to rear of DPOF; see also table 3A). Parameters in which the Machaerodus and Hans Johnson samples are smaller or shorter than, and beyond the S.D. range of, the XMas-Kat sample are no. 2 (palatal length), no. 6 (total cranial length, in part includes no. 2), and no. 36 (facial crest below DPOF; inversely correlated with no. 35). Parameters in which the Machaerodus and Hans Johnson samples are slightly shorter or lower (but at least partly within the S.D. range of) XMas-Kat are no. 5 (foramen magnum to rear of palate; combines nos. 3 and 4), no. 7 (premolar length; includes no. 2 in part), no. 8 (molar length; includes no. 2 in part), no. 9 (premolar-molar length; includes no. 2 in part), and no. 28 (orbital length). Parameters in which the Machaerodus and Hans Johnson samples are much greater or higher than the S.D. range of XMas-Kat are no. 35 (DPOF height; inversely correlated with no. 36; DPOF relative to facial crest). Those in which the above two samples are slightly higher than (but S.D. ranges overlap with) that from XMas-Kat are no. 18 (frontal width) and no. 29 (orbital height). In most cases where parameters differ, samples from Machaerodus and Hans Johnson quarries segregate together. In no. 33 (length, DPOF) and no. 38 (height of rear of DPOF relative to alveolar border), the Hans Johnson sample is partly (no. 33) to completely (no. 38) shorter than that from Machaerodus and XMas-Kat. In nos. 14–15, the S.D. range for the Machaerodus sample is overlapped by that from XMas-Kat, but the Hans Johnson sample overlaps that from XMas-Kat.

Except for parameter 2 (which likely is associated in part with nos. 7–9 and possibly with nos. 31, 33, and 34), no. 6 (in part includes no. 2), no. 36 (inversely correlated with no. 35), cranial parameters of the three quarry samples are generally similar. Thus, even though the three quarry samples are within about one mile of each other (Skinner and Johnson, 1984: fig. 16) and are nominally correlative, it appears that the Machaerodus and Hans Johnson samples have a slightly shorter face/palate and slightly taller DPOF than does the sample from XMas-Kat. As indicated in table 5, both sexes are represented in each of the three quarry samples, so these differences cannot be ascribed to sexual dimorphism. The possible significance of this is discussed below. In having the greatest number of specimens (table 5), the sample from XMas-Kat quarries is taken herein as the most representative of C. occidentale from Nebraska.

The upper incisors are well developed and ovate in cross section, with I3 tapering laterally. All have cement-filled infundibula. Lengths for I1–3 are 15.5 mm, 16 mm, and 15 mm, respectively; widths are 9.8 mm, 9 mm, and 8.4 mm. In F:AM 71800, the canine is relatively small, indicating the female sex of the specimen (in contrast to its much larger size in F:AM 71801; table 6Table 6 (Continued)). The canine is located 32 mm posterior to the rear of I3, or 36% of the length of the diastema (table 6Table 6 (Continued)) between I3 and P3. DP1 is absent in this as well as in all adult XMas-Kat quarry specimens. As indicated in table 6Table 6 (Continued), the canine is located about 34% of the length (mean) of the diastema in XMas-Kat and Hans Johnson material and about 31% of the length in Machaerodus Quarry specimens.

As considered from all quarry sites, the upper cheek tooth dentition is relatively complex. Protocones tend to be elongate, but only P3 and P4 have protocones with slightly lingually concave borders at this stage of wear. As indicated in table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }, there are two plis caballin in both the premolars and molars; fossette borders are relatively complex, except for the posterior border of the postfossette (usually with 2 plis only). Although some specimens of C. quinni in early wear stages show multiple plications on the anterior border of the prefossette (Woodburne, 1996b), this trait is especially characteristic of the C. occidentale group, with mean counts of three or more in most samples of P2–M1 (table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }). Overall, the sample from XMas-Kat Quarry has the most complex fossette borders at all tooth positions relative to the samples from the Machaerodus and Hans Johnson quarries. The fossette borders of P2, P3, and M2 appear to be more complex in the Machaerodus samples than in the Hans Johnson samples, with the reverse being the case for P4 and M1. M3 is not considered relevant in this regard as being poorly represented in the adult state.

The dP1 is absent from all adult skulls at the XMas-Kat quarries regardless of sex. The upper dentition of C. occidentale from the XMas-Kat quarries is relatively complex and high crowned (table 3B). As reconstructed from table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }, the unworn P2 is about 43 mm tall; P3 is 50 mm; P4 is 52 mm; M1 likely was about 66 mm; M2 was 65 mm; and M3 probably was about as tall as P2 (ca. 45 mm). At about medial wear (25–40 mm; table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }, excluding F:AM 71809, F:AM 71813), and as indicated in figures 9C and 10D, the mean plication count for P2 is about 6∶8∶6∶3. In P3, this is 5∶10∶7∶2; in P4, 7∶9∶7∶2; in M1, 4∶8∶7∶2; in M2, 4∶8∶6∶2; and in M3, 4∶7∶4∶1. For the entire sample, the protocone tends to be relatively elongate to subrounded (narrower in M2–3; table 3C), and the hypoconal groove is open and not spurred. The anterior border of the prefossette is relatively complex (more than 3 plis), commonly with many plications that are of small amplitude rather than there being a smaller number of relatively deep plis; the pli caballin is double or more in premolars and typically double in molars. Of eleven specimens, the P2 pre- and postfossette score is confluent in three specimens, separate in four specimens, and not applicable in four specimens.

From Machaerodus Quarry

The dP1 is absent in all (two) adult male skulls and one adult female. It is present in one adult female skull. The upper dentition of C. occidentale from the Machaerodus Quarry is somewhat smaller than the sample from XMas-Kat Quarry and slightly less complex. The unworn cheek tooth crown height is difficult to assess in the available sample. Based on F:AM 71844, P3 in an adult stage of wear is about 40 mm tall (table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }). M3 (very early wear) is 41 mm tall in F:AM 71835 (table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }), so the unworn crown height would have been at least 43 mm. This is comparable to the MSTHT in the XMas-Kat and Hans Johnson samples, so it is likely that the other cheek teeth were about as tall as in those samples. In about medial wear (ca. 30 mm; F:AM 71844, F:AM 71835, table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }), complexity in P2 reaches a mean of 5∶6∶3∶1; in P3, 5∶7∶7∶1; in P4, 3∶8∶6∶2; in M1, 2∶7∶7∶2; in M2, 3∶8∶6∶2; and in M3, 1∶7∶5∶1. With regard to the entire sample, the protocone tends to be elongate in the premolars but somewhat more ovate in the molars; the protocone retains a slight lingual concavity into later wear; the hypoconal groove is about equally spurred as un-spurred and open; the anterior border of the prefossette tends to be relatively complex and well delineated in medial wear, especially in the premolars; and the pli caballin is typically double in both premolars and molars but tends to be somewhat more complex in the premolars. Overall, the upper cheek teeth are less complex than in the XMas-Kat Quarry sample. The sample apparently is about as hypsodont as that in the Hans Johnson Quarry. Of five specimens, the pre- and postfossette of P2 score confluent in two, separate in one, and not applicable in two.

From Hans Johnson Quarry

The dP1 is absent in all three adult male skulls and is present in all four adult female skulls. The upper dentition of C. occidentale from the Hans Johnson Quarry is relatively complex and high crowned. The unworn P2 is about 45 mm tall, judging from the very early wear stage of F:AM 71862 (table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }), and P3 is likely about 50 mm tall, judging from the very early wear in F:AM 71862 and compared with materials from the XMas-Kat quarries. On the same basis, likely unworn crown heights for P4 are 52 mm; M1 likely was about 64 mm; M2 was about 65 mm; and M3 probably was about as tall as P2 (ca. 45 mm). At about medial wear (25–40 mm; F:AM 71862, F:AM 71861, F:AM 71860, table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }), the plication count for P2 is about 4∶5∶5∶1; for P3, 3∶6∶5∶1; for P4, 5∶6∶5∶1; for M1, 3∶8∶5∶2; and for M2, 2∶5∶5∶1. M3 is not represented at this wear interval. In the total sample, the protocone tends to be relatively elongate (table 3C) and lingually concave; the hypoconal groove is open and not spurred. The anterior border of the prefossette is relatively complex (3 or more plis in P2–M1), usually being represented by relatively deep plications rather than by a larger number of plications having a very shallow amplitude, and the pli caballin is double or more in premolars, typically double in molars. Overall, the dentition of the Hans Johnson sample is most similar to that of the type of C. occidentale than any other sample under study here. This similarity includes, in F:AM 71861 and F:AM 71860, the disconnection of the pli caballin from the hypoconal selene, with that selene penetrating labially to incorporate or lie in proximity with the pli protoconule, also as seen in P3 of the holotype. Of seven specimens, the pre- and postfossette in P2 score confluent in three, separate in three, and not applicable in one.

From Ed Ross Ranch Quarry, South Dakota

F:AM 71872 from ?late medial Clarendonian sites (referred by Skinner and MacFadden to Hollow Horn Bear Quarry [sic]), apparently in the Thin Elk Formation (♂ skull with right and left C1, P2–M3; fig. 11; tables 3A, 4C, 5), also indicates the presence of taxa with the morphology of C. occidentale in this district of South Dakota. The unworn MSTHT likely was in the range of 60 mm (table 4C). The cranial morphology shows that the ratio of facial length of the lacrimal (nos. 32, 39, table 5) to the length of the POB (no. 32, table 5) is relatively large (0.75; table 3A), and the ratio of the length of the lacrimal tip to the rear DPOF (no. 39, table 5) to the POB length is generally comparable (0.25; table 3A) to the later Clarendonian taxa C. occidentale or C. skinneri, n.sp. AMNH 10869, also possibly from the Thin Elk Formation, has an unworn MSTHT (M1) of about 65 mm (table 4C). As indicated above, this material is a likely correlate of that from the XMas-Kat quarries. P2 pre- and postfossettes are confluent on the right side of F:AM 71872 and separate on its left side; in AMNH 10869, the condition is separate.

The mandible of C. occidentale at the XMas-Kat Quarry is represented by F:AM 71800. As shown in MacFadden (1984: fig. 135) and table 8 (nos. 10–12), this female mandible is relatively shallow below the cheek tooth row. The symphysis extends for about half the length of the diastema from p2 to c1. The mandibular foramen is sited about halfway in the diastemal region. The lower incisors are arranged in an arcuate fashion and display cement-filled infundibula. The lower canine is nearly directly posterior to i3. The lower cheek tooth dentition is relatively slender in overall proportions. Protostylids are present on all cheek teeth except p2 and take the shape of a labially projected enamel loop. The metaconids and metastylids are well developed, of equal size, and essentially subrounded, except in p2, where the metaconid is slightly smaller than the metastylid and subtriangular. C. occidentale differs from C. quinni in the clear separation of p2 metaconid and metastylid and in their larger and subequal size, whereas in C. quinni the metaconid tends to be larger than the metastylid and both cuspids are relatively smaller. The labial margins of the protoconid and hypoconid are relatively flat in p2 and p3 but more convex in the remaining cheek teeth. A single pli caballinid is present in each of the premolars but not in the molars. The ectoflexids are labial in position in all cheek teeth and do not penetrate the metaconid–metastylid isthmus. The linguaflexid is relatively shallow in premolars and deeper in molars, but still sub V-shaped rather than U-shaped.

F:AM 71844, from the Machaerodus Quarry, is similar overall to F:AM 71800. The diastema in this male mandible is slightly shorter than in F:AM 71800, but other dimensions are comparable (table 8). The rear of the symphysis is about level with the anterior border of the mandibular foramen instead of being located posterior to the foramen in F:AM 71800. F:AM 71844 is in a slightly earlier wear stage than F:AM 71800, as judged by the less worn heel of m3. This may account for the otherwise unusual penetration of the isthmus by the ectoflexid in p2, p3, and m1–3 in the Machaerodus specimen and the overall less prominently developed metaconids and metastylids. Woodburne (2003: 402) indicates that the penetration of lower premolar isthmuses by the ectoflexid is a plesiomorphic trait that is found in Merychippus insignis but not, for example, in C. goorisi. As indicated in the present report, this trait is typically absent in juvenile or adult (but not ancient) specimens of other species of Cormohipparion, so its presence in F:AM 71844 apparently is an atavistic occurrence.

F:AM 71855, from the Hans Johnson Quarry, is a male in late wear. The mandible is somewhat crushed dorsoventrally, but its overall dimensions (table 8) appear comparable to those of the other two specimens. As in F:AM 71844, the symphysis ends anterior to the mandibular foramen, which is located near the mid-length of the diastema. The cheek teeth in F:AM 71855 show that the ectoflexids penetrate the isthmus in late wear on all cheek teeth. Ectoflexids are present in all teeth except p2 but are nearly obscured by wear in p3–m1.

Cormohipparion matthewi, new species

figures 14–15; tables 2–3Table 3 (Continued)Table 3 (Continued), 6Table 6 (Continued), 8–Table 910

Figure 14

Cormohipparion matthewi. F:AM 71802, XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska. A, lateral view of cranium. B, occlusal view of upper cheek tooth dentition.

Table 9

Measurements (mm) of Cranial Parameters for Cormohipparion matthewi, Merritt Dam Member, Ash Hollow Formation, medial Clarendonian Nebraska

Type Specimen

F:AM 71802, nearly complete skull with right and left I1–3, C, P2–M3, female, late adult wear.

Type Locality

XMas-Kat quarries, Merritt Dam Member, Ash Hollow Formation, Cherry County, Nebraska.

Age

Medial Clarendonian.

Distribution

Xmas-Kat quarries, north-central Nebraska.

Description of Type Specimen

The cranium of F:AM 71802 (fig. 14) generally resembles that of C. occidentale (fig. 10). It is slightly crushed dorsoventrally and medially above the facial crest. The frontals are slightly domed above the dorsal skull profile above the nasal bones. The canine is situated about two-thirds of the length of the diastema anterior to P2. The DPOF is relatively small (short in length as well as height; nos. 33 and 35, table 9) and appears to have been pocketed to a depth of about 13 mm posterior to the rear of the DPOF and, thus, anterior to the tip of the lacrimal. The IOF is located above the anterior half of P3 and below the anterior end of the DPOF, as in Cormohipparion generally. The anterior end of the orbit is located above M3 rather than posterior to it. The lacrimal is blunt anteriorly, with its anterior end being about aligned with the maxillo-jugal suture rather than extending anteriorly from that point. It extends across about 60% of the POB (length 2, table 3A). The anterior tip of the lacrimal is about 20 mm posterior to the rear of the DPOF so that the facial extent of the lacrimal is about 27.5 mm, or 40%, of the DPOF (length 1, table 3A). DP1 appears to be absent. As indicated in table 10, P2 is at least 35 mm tall, and P3 is at least 33 mm tall. Based on the hypodigm, the cheek teeth appear to have been 40–45 mm tall in the unworn condition (table 3B), in contrast to 50–66 mm in C. occidentale. The fossette plications appear to have been relatively complex, as discussed below, but not as complex as in C. occidentale.

Diagnosis

Based on the type and referred material, and as indicated in figure 15 and in comparison with figure 13, C. matthewi differs from C. occidentale in the greater length of the postpalatal dimension of the cranium (no. 3), and thus somewhat greater length of no. 5 (which includes no. 3). Effectively, all other measured parameters in C. matthewi are smaller, narrower, or shorter than in C. occidentale. This is exemplified by the shorter muzzle (no. 1) in C. matthewi. The palatal length (no. 2) and total cranial length (no. 6) in C. matthewi is also smaller than in C. occidentale from XMas-Kat Quarry but not in the samples from the Machaerodus and Hans Johnson quarries. Similarly, the cheek-tooth lengths (nos. 7–9) in C. matthewi are at the lower end of the range for these features in the XMas-Kat Quarry sample of C. occidentale but more in line with those from the Machaerodus and Hans Johnson quarries. The snout in C. matthewi is much narrower (no. 14) than in C. occidentale and all other species of the C. occidentale group except C. johnsoni, n.sp., which is even narrower. C. matthewi has the narrowest transglenoid dimension (no. 19) of any species discussed here. The snout (no. 25) is relatively much shorter vertically in C. matthewi, C. johnsoni, n.sp., and C. fricki, n.sp., than in C. occidentale or the other species. The nasal notch is relatively less developed in C. matthewi, C. johnsoni, n.sp., C. merriami, n.sp., C. fricki, n.sp., and C. skinneri, n.sp., compared with C. occidentale. The DPOF (no. 33) is sharply shorter in C. matthewi and C. johnsoni, n.sp., and the S.D. ranges are shorter in C. fricki, n.sp., and C. merriami, n.sp. The separation of the IOF relative to the rear of the DPOF (no. 34) is much less in C. matthewi compared with C. occidentale or the other species, except C. johnsoni, n.sp., and the DPOF is much shorter vertically (no. 35) in C. matthewi than in C. occidentale. The rear of the DPOF is nearer the alveolar border (no. 38) in C. matthewi than in C. occidentale and is exceeded in this only by C. johnsoni. C. matthewi has a much shallower DPOF (no. 40) than any other species.

The unworn dentition is not represented; overall plications are less numerous in medial adult wear; and protocones are perhaps somewhat more elongate, but the sample is small. The unworn MSTHT for the upper cheek teeth is estimated at 40–45 mm (tables 3B, 10).

Referred Material

From XMas-Kat quarries, eastern Cherry County, Nebraska (Skinner and Johnson, 1984: 315, fig. 16). F:AM 71804, nearly complete skull, RI3, right and left P2–M3, female, adult wear. F:AM 71805, partial cranium, RI2–3, right and left C, P2–M3, male, early adult wear.

From Machaerodus Quarry, eastern Cherry County, Nebraska (Skinner and Johnson, 1984: 315, fig. 16). F:AM 71833, juvenile cranium, with LI3, LdP1, right and left dP2–4, M1, M2 erupting.

From Hans Johnson Quarry, eastern Cherry County, Nebraska (Skinner and Johnson, 1984: 315, fig. 16). F:AM 71859, nearly complete ♂ skull; snout displaced; dorsoventrally somewhat flattened, laterally expanded, right and left I1–2, LC, right and left P2–M3. F:AM 71863, nearly complete ♀ cranium of immature individual, LI1–3, R, C erupting, right and left dP1–dP4, M1 erupting.

Description of Referred Material

F:AM 71895 and F:AM 71863 are too crushed to yield significant new morphological information, but F:AM 71804 and F:AM 71805 confirm that the blunt end of the lacrimal is effectively aligned with the nearly vertical maxillo-jugal suture and ends about 18 mm and 13 mm, respectively, posterior to the rear rim of the DPOF (no. 39, table 9). F:AM 71863 is somewhat crushed in this region, but its right side confirms that the lacrimal terminates about 23 mm posterior to the rim of the DPOF. F:AM 71833 is a juvenile cranium and yields no significant information.

DP1 appears to be absent, except in the juvenile cranium F:AM 71833, where the tooth is very well worn and about 9 mm long, and in the equally juvenile cranium F:AM 71863 (female), where the tooth is also worn and only about 5 mm long. DP1 is absent in all adult crania, regardless of sex. Compared with the average length of P2 in the adult specimens from the XMas-Kat quarries (28.6 mm; table 10), the dP1–P2 length ratio would be 0.32–0.18, generally consistent with the late medial Clarendonian age of this sample. The cheek teeth appear to have been 40–45 mm tall in the unworn condition, in contrast to 50–66 mm in C. occidentale (tables 3B, 10). As indicated in table 10, the anterior border of the prefossette in C. matthewi has a mean of at least 3 plis in P2 (mean of 6), P3, M2, and M3 (mean of 5), but a mean of 2 in P4 and M1. For P2–M2, the posterior border of the prefossette usually has a mean of at least 7 (up to 8) plis, with a mean of 2 in M3. The anterior border of the postfossette usually has a mean of 6 plis in P2, P3, and M1, but a mean of 5, 4, and 2 in M2, P4, and M3. The posterior border of the postfossette has a mean of 2 plis in P2 and P3 but a mean of 1 pli in the other cheek teeth. Regarding the confluence of pre- and postfossettes of P2, the three XMas-Kat specimens are scored as one yes, one no, and one not applicable; for the Machaerodus Quarry, the score is one not applicable; for the Hans Johnson Quarry, the score is one not applicable and one no. Based on the very small sample, it appears that the pre- and postfossettes in P2 may be confluent somewhat less than 50% of the time.

At present, no mandibles or lower dentitions can be directly associated with crania referred to C. matthewi.

Cormohipparion johnsoni, new species

figures 16–Figure 1718; tables 2–3Table 3 (Continued)Table 3 (Continued), 6Table 6 (Continued), 11–Table 11(Continued)12Table 12 (Continued)

Table 11

Measurements (mm) of Cranial Parameters for Cormohipparion johnsoni and C. merriami, Burge Member, Valentine Formation, early Clarendonian Nebraska

Table 11(Continued)

Type Specimen

F:AM. 71891, ♂ skull, missing snout and occiput, somewhat crushed dorsoventrally and laterally skewed, dentition in very late wear.

Type Locality

Burge Quarry, Burge Member, Valentine Formation, Nebraska.

Age

Early Clarendonian, Burge Local Fauna, Nebraska.

Distribution

Type locality.

Description of Type Specimen

The basic morphology of the type cranium is as for C. occidentale. The left side best preserves the morphology in lateral view (fig. 16). As indicated in table 11Table 11(Continued), and in comparison with the juvenile cranium F:AM 71894 (fig. 17), the orbit is somewhat attenuated anterodorsally, and its anteroventral portion faces more posterolaterally than otherwise normal. The anteroventral margin of the orbit is less sharply demarcated than in F:AM 71894, for instance. The frontal region of the cranium is depressed in the type, and the postorbital fossa is distorted due to the adjacent lateral wall of the cranium having been displaced somewhat ventrally and laterally. As indicated in figure 16, the lacrimal exits the orbit just anterior to the end of the rugose dorsal orbital border and extends anterodorsally to join the fronto-nasal suture, then descends shallowly anteroventrally toward the rear of the DPOF. The lacrimal extends anteriorly about to the maxillo-jugal suture and nearly reaches, but does not penetrate, the rear of the DPOF. The lacrimo-jugal suture descends posteroventrally to enter the orbit well below the lacrimal foramen. As indicated in table 11Table 11(Continued), the POB is about 40 mm long. The DPOF is well defined posteriorly. Its dorsal margin extends anteroventrally to the relatively well-defined anterior end, situated above and in close proximity to the IOF. The bony surface of the DPOF is broken so that the canal for the maxillary nerve is exposed posterior to the infraorbital foramen, but the continuous ventral margin of the DPOF can be recognized as it joins the posterior margin of the fossa. The IOF is located about 44 mm (table 11Table 11(Continued)) above the rear half (metacone) of P3 and is aligned about with the orbital bisector. Allowing for distortion, it appears that the DPOF is positioned relatively high on the face, anterior to the orbit in C. johnsoni.

In lateral view (fig. 16A), the robust canine of this male cranium can be seen, but the snout anterior thereto is not preserved. As indicated in table 6Table 6 (Continued), it appears that the diastema would have been as extensive as in F:AM 125899 but probably more robust, reflective of its male sex. The basic morphology is similar to that of C. occidentale. Roots for dP1 are preserved, but the crown is absent. Due to the advanced age of the specimen, the dental morphology is poorly preserved. That for P2 is virtually obliterated, with the protocone confluent with the protoloph at this advanced stage of wear (MSTHT 10 mm, table 12Table 12 (Continued)). In P3, the protocone is confluent to still virtually isolated (left side; fig. 16B); it is barely open to the protoloph in M1 but isolated in all other teeth. Tables 3B and 12Table 6 (Continued) show that the protocone is ovate in all teeth except for M3, where it is more elongate (protocone width to length ratio of 0.47), with M3 preserving a remnant of what likely was a fairly complex fossette border morphology throughout. Due to the late stage of wear shown by this specimen, the teeth are very low crowned. Cranial dimensions are summarized in table 11Table 11(Continued).

Diagnosis

Based on the type and referred material, Cormohipparion johnsoni is distinguished from all other species of the subgenus Cormohipparion in having an unworn mesostyle crown height of 40–50 mm (table 3B). In addition, figure 18 in comparison with figure 13 indicates that C. johnsoni differs from C. occidentale and the other species in having a much shorter muzzle (no. 1) and narrower muzzle at the I3 (no. 15). C. johnsoni differs from the XMas-Kat sample of C. occidentale, but from not the Machaerodus and Hans Johnson samples, in the shortness (no. 2) and lesser palatal breadth (no. 13) and in having shorter premolars (nos. 7, 9). It differs from all samples of C. occidentale in the diminished breadth of the frontals (no. 18) and transglenoid region (no. 19) and from all other species in these parameters, as well (except for no. 19 in C. matthewi). C. johnsoni also has smaller orbit (nos. 28–29) than in any other species of the group, as perhaps befits its plesiomorphic status (see below). The POB is shorter in C. johnsoni and C. merriami, n.sp., than in the other species. The DPOF (no. 33) is sharply shorter in C. johnsoni (and C. matthewi) and ranges to a comparably small size in C. merriami, n.sp., and C. fricki, n.sp. The separation of the IOF relative to the rear of the DPOF (no. 34) is much less in C. johnsoni as compared with C. occidentale or the other species, except C. matthewi, and the lacrimal nearly reaches the rear of the DPOF (no. 39) and extends across about 80% of the POB (table 3A), in contrast to any other species.

Especially as regards parameters 2 and 39, the very large range for the standard deviation is influenced by the small number of specimens, and the overall smaller dimensions can be related to the overall smaller size of the skull of C. johnsoni compared with C. occidentale. Still, the very diminished size of parameters 14 (narrow diastema) and 39 (tip to rear of DPOF; table 11Table 11(Continued)) is noteworthy, as is the fact that other parameters are near or contained within the range of those in C. occidentale (fig. 18): no. 7, premolar row length; no. 8, molar row length; no. 12, choanal width; no. 31, facial length; no. 35, height of DPOF relative to the facial crest; no. 36, relative position of the facial crest and DPOF; no. 37, relative height of the IOF from the alveolar row; no. 38, height of rear of DPOF relative to alveolar row; and no. 40, medial depth of DPOF.

Cormohipparion johnsoni (fig. 18) differs from C. matthewi (fig. 15) in having a much shorter snout and palate (nos. 1–3) in contrast to comparably sized premolar and molar rows (nos. 7–9). C. johnsoni also has a narrower palate (no. 13); a greatly narrower diastemal region (no. 14) and muzzle (no. 15); a much narrower frontal width (no. 18); and a narrower transglenoid width (no. 19). In contrast to an equally high snout (no. 25) and equally incised nasal notch (no. 30) relative to C. matthewi, C. johnsoni has a smaller orbit (nos. 28–29), a shorter facial region (no. 31), a relatively shorter POB (no. 32), and a somewhat shorter DPOF (no. 33). However, the IOF location relative to the DPOF is similar (no. 34) in C. johnsoni and C. matthewi, but C. johnsoni has a taller DPOF (no. 35), a diminished separation of the facial crest from the DPOF (no. 36), a decreased separation between the IOF and alveolar border (no. 37), a greater separation between the DPOF and alveolar border (no. 38), a lacrimal that is very close to the rear of the DPOF (no. 39; table 11Table 11(Continued)), and a medially deeper DPOF (no. 40).

In comparison with C. merriami, n.sp. (fig. 20), C. johnsoni (fig. 18) has a much shorter snout and palate (nos. 1–2); a much shorter premolar and molar row (nos. 7–9); a much narrower palate (no. 13), diastemal region (no. 14), muzzle (no. 15), frontal region (no. 18) and transglenoid region (no. 19); a much smaller snout (no. 25); a smaller orbit (no. 29); a shallower nasal notch (no. 30); and a much shorter facial region (no. 31). Whereas the POB is comparable (no. 32) in the two species, C. johnsoni has a much shorter DPOF (no. 33); a decreased separation between the IOF and rear of the DPOF (no. 34); a much reduced DPOF height (no. 35); a somewhat diminished separation of the facial crest and DPOF (no. 36); and of the IOF relative to the alveolar border (no. 37), a more normative (cf. C. occidentale) separation of the alveolar border and rear of DPOF (no. 38); a lacrimal anterior tip very close to the rear of the DPOF (no. 39, table 11Table 11(Continued)); and a comparable DPOF medial depth.

In addition to overall size and crown height, as well as dental complexity, C. johnsoni differs from C. quinni in the more anterior position of the DPOF, reflected in the fact that the lacrimal barely reaches, but does not penetrate, the DPOF. C. johnsoni is the most plesiomorphic member of the C. occidentale group.

Referred Material

F:AM 71894, nearly complete immature ?♀ cranium, with dP1–4 present and well worn. F:AM 125899, laterally compressed cranium with occipital region obscured, right and left C, dP1–M2; M3 not erupted. F:AM 119095, right maxilla with P3–M3.

Description of Referred Material

F:AM 71894 (fig. 17) is useful in demonstrating the presence of dP1 as a relatively small, but still functional, tooth (table 12Table 12 (Continued)). The deciduous premolars are in an early stage of wear, with the pattern not fully developed on dP2 or dP4. The dP3 and dP4 show double plis caballin and complex opposing borders of the pre- and postfossette. F:AM 125899 indicates that the ratio of the length dP1 to the length of P2 is 0.3 (table 12Table 12 (Continued)).

In lateral view (fig. 17), the well-developed and strongly demarcated DPOF is clearly visible, as is the wide POB and the pointed lacrimal that terminates a short distance posterior to the rear of the DPOF (5.5 mm, no. 39, table 11Table 11(Continued)). Although not as well shown, a similar arrangement for the lacrimal is interpreted to have obtained in the type and in F:AM 125899, with the anterior tip of the lacrimal 11 mm posterior to (type) or 6.5 mm to the rear of the DPOF (F:AM 125899).

F:AM 125899 and F:AM 119095 indicate that the unworn cheek tooth MSTHT dimensions apparently were about 45–50 mm (tables 3B, 12Table 12 (Continued)). The cheek teeth of F:AM 125899 were removed from the cranium under the direction of the late Morris F. Skinner, and measurements were taken before they were sectioned at three intervals at progressively deeper levels in each tooth. The tooth segments were preserved by being glued to a plate of glass. From these segments it can be seen that M3 is completely unworn and is not furnished with a cement cover, as are the other teeth. The MSTHT for M3 is not available (encased in matrix). P4, the next tooth in the wear progression, shows only incipient wear on its highest crests and is 48.6 mm tall. In further order of increasing wear, P2 MSTHT is 36.7 mm; P3 is 42.2 mm; M2 is 45.6 mm; and M1, with a nearly complete wear pattern, is 42.4 mm tall. It is likely that M1 originally was about as tall as P4. Based on its currently unworn state, the MSTHT for P4 appears to have been no greater than 50 mm. Based on F:AM 125899, the unworn MSTHT for upper cheek teeth in C. johnsoni appears to have been 40–50 mm. These parameters are basically verified by AMNH 119095 (table 12Table 12 (Continued)), which also shows the moderate complexity of the pre- and postfossettes. In P3–M3, the anterior border of the prefossette and posterior border of the postfossette typically have a single pli (mean value, table 12Table 12 (Continued)). In P3, opposing borders of the pre- and postfossette have mean values of 5 and 4 plis, respectively. In P4, comparable values are 6 and 5; in M1 and M2, 6 and 6; and in M3, the values are 7 and 3. The protocones are ovate to slightly elongate and ovate, with somewhat flattened lingual borders in P3 and P4 but lingually rounded in the molars. As indicated in table 3C, the mean width-to-length ratios for the cheek teeth of C. johnsoni are generally similar to those for C. merriami, n.sp., and generally wider (at or above 50% in P2–M2) than in all other species of the C. occidentale group. AMNH 119095 displays a pli protoconule that is isolated from the posterolingual border of the prefossette in early wear of P3–4 and a large and distinct loop in this position in M1–3. The pre-and postfossettes were separated from one another in P2 of F:AM 125899. Plis caballin are more complex (mean value of 2, table 12Table 12 (Continued)) in the P3 and P4 compared with the other cheek teeth. The hypoconal groove is usually simple (bears a spur in the juvenile P2 and P3 in F:AM 71984). In early wear (F:AM 125899), P2 shows separated pre- and postfossettes, but in late wear (type specimen) these fossettes are confluent.

At present, no mandibles or lower dentitions can be directly associated with crania referred to C. johnsoni.

Discussion

As discussed below, C. johnsoni is the basal member of the C. occidentale group.

Cormohipparion merriami, new species

figures 19–20; tables 2–3Table 3 (Continued)Table 3 (Continued), 6Table 6 (Continued), 11Table 11(Continued), 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] }

Neohipparion cf. coloradense Webb 1969: 98.

Type Specimen

AMNH 141219, nearly perfect ♀ skull, right and left I3, C, dP1–M3 in about mid wear.

Type Locality

June Quarry, Burge Member, Valentine Formation, Nebraska.

Age

Early Clarendonian, Burge Local Fauna, Nebraska.

Distribution

Midway Quarry, Burge Member, Valentine Formation, Nebraska.

Description of Type Specimen

AMNH 141219 is generally similar to C. occidentale in basic skull morphology (fig. 19A). The IOF is situated above the anterior half of P3 and below the anterior end of the DPOF, as is typical of the C. occidentale group of Cormohipparion. The anterior edge of the orbit is located above M3 rather than more posteriorly, as in C. occidentale. The lacrimal is pointed anteriorly and extends about 11 mm anteriorly beyond the maxillo-jugal suture but ends about 15 mm (mean value) posterior to (and does not reach) the rear of the DPOF (no. 39, table 11Table 11(Continued)). It appears that the DPOF is pocketed posteriorly to about the anterior tip of the lacrimal. In contrast to C. occidentale as represented by the XMas-Kat sample, the MSTHT is estimated as having been ca. 55 mm (i.e., much lower crowned; tables 3B, 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] }); dP1 is present (the ratio of the length of dP1 to the length of P2 is 0.37) in the type specimen but apparently has been lost in the juvenile cranium F:AM 71893. As indicated in table 3C, the protocones are relatively wider than in all members of the C. occidentale group except for C. johnsoni. Whereas P2 is subovate, the protocone in the other cheek teeth is more elongate (fig. 19B). Fossette plications are discussed in conjunction with the referred material (see below).

Diagnosis

Based on the type and referred material, Cormohipparion merriami is distinguished from other species of the subgenus Cormohipparion in having lacrimal that extends anteriorly beyond the maxillo-jugal suture but still does not reach the rear of the DPOF. The MSTHT is relatively low (ca. 55 mm) in C. merriami; the protocones are relatively wide and not so elongate. As indicated in figure 20 and in comparison with figure 13, C. merriami is greater than (or likely would range much higher than) all other species in length of palate (no. 2) and has a relatively greater premolar (no. 7) and thus cheek tooth (no. 9) length than in C. occidentale (unknown in C. johnsoni), with the exception of the palate and cheek teeth of C. skinneri, n.sp. Cormohipparion merriami tends to have a smaller orbit than in C. occidentale (nos. 28–29) and all other species, except for C. johnsoni, in which the orbit is still smaller. The nasal notch in C. merriami is less strongly incised than in C. occidentale (no. 30) but more so than in all other species. The facial length (no. 31) apparently is (or likely would range) larger than in C. occidentale and all other species except C. skinneri, n.sp. The POB (no. 32) tends to be shorter in C. merriami than in C. occidentale and all species except the comparably sized C. johnsoni.

Referred Material

From June Quarry, Burge Member, Valentine Formation, Nebraska. AMNH 141273, left maxillary fragment with dP1–dP3; AMNH 141274, right upper cheek tooth row in very early wear, P2–M3 and LM2; AMNH 141272, right upper cheek tooth row in medial wear, P3–M3.

From Midway Quarry, Burge Member, Valentine Formation, Nebraska. AMNH 141220, old adult, cracked and crushed skull with relatively intact facial region; F:AM 71893, juvenile partial cranium LI1–2, right and left I3, alveoli for the canines, right and left dP2–4 erupted and worn to the degree that the enamel pattern is almost completely developed; right and left M1 are barely exposed above the gum line. The cranium is largely undistorted but lacks the occipital region and paroccipital processes. F:AM 71983 also shows the triangular lacrimal bone seen in the type and is similar in that the lacrimal does not reach the rear of the DPOF.

Description of Referred Material

The partial cranium AMNH 141220 retains the muzzle but lacks the intervening nasal elements to the cheek teeth, and the skull itself is conspicuously cracked, with most of the left side of the facial region missing. Overall, AMNH 141220 appears similar to the type skull. The POB is intact and shows the lacrimal as about subtriangular in aspect, with its anterior end about 17 mm posterior to the rear of the DPOF (no. 39, table 11Table 11(Continued)), comparable to that of the type specimen. The DPOF in AMNH 141220 is somewhat longer than in the type and somewhat shorter dorsoventrally but otherwise similar, with the differences (including the apparently greater DPOF-IOF distance, no. 34) due largely to the cracked and distorted condition in AMNH 141220.

The upper cheek tooth dentition displayed by the referred material includes the presence of dP1 in AMNH 141273, in association with a strongly worn dP2–3, so that its persistence in the adult condition as shown by the type specimen is to be expected. In AMNH 141273, dP1 is 9.8 mm long and 6 mm wide, somewhat smaller than the dimensions (12.1 mm and 6.7 mm) in the type. As indicated in table 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] }, the ratio of the length of dP1 to the length of P2 is 0.37 in the type. AMNH 141274 is the least worn of the available specimens and suggests that the unworn MSTHT for the cheek tooth dentition was likely 45–55 mm tall, with P4 and M1 having been 52–55 mm tall (table 3Table 3 (Continued)Table 3 (Continued)). The upper cheek tooth protocones were uniformly elongate and ovate, except in P3 and P4, where the lingual border is essentially flat (or slightly concave) and, in late wear of P2, where the protocone is connected to the protoloph in AMNH 141220. As indicated in tables 3C and 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] }, the mean width-to-length ratios are 0.67, 0.50, 0.47, 0.51, 0.53, and 0.46 for P2–M3. The type (left side, but not the right side) and AMNH 141274 show that the pre- and postfossettes are confluent labially in P2 in relatively early wear (16% and 25% wear, respectively), but in late wear (AMNH 141220) these fossettes are separate.

Tables 2 and 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] } show that the fossette borders overall are simpler than in C. occidentale or C. matthewi. The mean values in table 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] } indicate that C. merriami has a more complex plication pattern than C. johnsoni (table 12Table 12 (Continued)). In table 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] }, it can be seen that the anterior border of the prefossette in P2 has a mean value of 5 plis, but in all other cheek teeth this is reduced to 2 plis or 1 pli (M1). In premolars, the posterior border of the prefossette contains a mean value of 5–10 plis; in the molars, this feature is represented by 6–9 plis. Remaining mean values for fossette plications are anterior border of the postfossette 4–6 in premolars and 3–7 in molars. The mean value of the plis in the posterior border of the postfossette is a single pli in all cheek teeth but M2 (where the mean value is 2).

Discussion

Webb (1969) discussed the morphology of a sample from Burge Quarry, Valentine Formation, Nebraska, allocated to Neohipparion cf. coloradense. MacFadden (1984) reviewed the species referable to Neohipparion and defined N. coloradense as having an MSTHT of 35–43 mm, a simple cheek tooth occlusal pattern, and a relatively ovate protocone. These aspects of Webb's Burge sample (MSTHT 47–52 mm; cheek teeth with relatively complex occlusal pattern, and a more elongate protocone) are most similar to C. merriami as herein defined and characterized.

C. merriami is comparable to C. johnsoni in retaining dP1 into the adult condition but more advanced in being taller crowned and in the greater complexity of the fossette borders of the upper cheek teeth. It also is of larger size than C. johnsoni, with which it was contemporaneous. See below for further discussion of C. merriami relative to the other species of the C. occidentale group.

At present, no mandibles or lower dentitions can be directly associated with crania referred to C. merriami.

Cormohipparion fricki, new species

figures 8, 21–22; tables 2–3Table 3 (Continued)Table 3 (Continued), 6Table 6 (Continued), 14–15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }

Figure 21

Cormohipparion fricki, F:AM 73912, MacAdams Quarry, Clarendon beds, Texas, medial Clarendonian. A, lateral view of craniuim. B, occlusal view of upper cheek tooth dentition.

Table 14

Measurements (mm) of Cranial Parameters for Cormohipparion fricki, Medial Clarendonian, Texas and South Dakota Characters are illustrated in figures 3–Figure 45.

Type Specimen

F:AM 73912, partial skull with right and left C–M3, adult male, M3 well worn.

Type Locality

MacAdams Quarry, Clarendon beds, Texas.

Age

Medial Clarendonian, Nebraska, Texas, South Dakota.

Distribution

Type locality and Hollow Horn Bear Quarry, Ash Hollow Formation undifferentiated (Skinner and Johnson, 1984; figs. 2A, 3; table 1), Todd County, South Dakota.

Description of Type Specimen

As indicated in figure 21, the specimen is relatively complete. The nasal bones above the snout are missing, as are the upper borders of the orbits. The cranium is slightly depressed by postmortem action, so that certain cranial width dimensions are exaggerated. In comparison, F:AM 71800 from the XMas-Kat sample is somewhat less distorted. The upper cheek tooth dentition is well preserved. The overall quality of F:AM 73912 qualifies its being chosen for the type specimen, even though some others (e.g., F:AM 73920) better preserve the incisors.

In overall morphology, the cranium of C. fricki is comparable to that of many contemporaneous equids. The nasal notch appears to have been located about midway between the canine and the anterior tip of P2, but the relatively elongate facial region as compared with, for example, C. quinni (Woodburne, 1996b), is shown by the anterior margin of the orbit being located above the rear portion of M3 in adult wear. In lateral view (fig. 21A), the profile overall is relatively regular, with the dorsal surface (lacking the nasal bones) being gently curved. Although depressed postmortem, the frontal bones do not appear to have been dome-like. The orbits are relatively large (nos. 28–29; table 14) and the POB (no. 32) is relatively wide. The DPOF is thus well anterior to the orbit and is higher posteriorly than anteriorly but not as distinctly teardrop shaped as, for instance, in the XMas-Kat taxon. As in Cormohipparion generally, the IOF is located just below the anteroventral margin of the DPOF, slightly posterior to the actual anterior end of the fossa, as is the usual case. The anterior end of the DPOF is not as pronounced as typically seen in XMas-Kat materials of Cormohipparion. The anteriorly wide POB is reflected in the lacrimal usually being exposed on the posterior one-half of the preorbital bar (length 2, table 3A; no. 39, table 14) and not entering the rear of the DPOF, as in C. quinni. In the type, dorsoventral compression of the facial region results in the ventral boundary of the lacrimal being distorted. Based on other specimens, the morphology of the lacrimal would resemble that in F:AM 71800 (fig. 8A), further described below, rather than being narrowed anteriorly. As in C. quinni and the XMas-Kat materials, the facial crest ends about in alignment with the premolar/molar boundary at an elevation approximately midway between the alveolar border and the ventral margin of the DPOF. In spite of the wider POB, the DPOF is still situated about as in C. quinni; the rear of the DPOF is still located about above the mesostyle of M1. In the type and other specimens of C. fricki, the lacrimal is approximately as long as in C. quinni, so that the extension of the POB appears to have taken place between the lacrimal tip and the rear of the POB. This is reflected in dimension no. 32 being essentially twice that in C. quinni (table 14, in comparison with table 16, no. 32, in Woodburne, 1996b), even though the facial length of the lacrimal is about 34 mm in C. fricki versus about 25 mm in C. quinni.

Table 16

Measurements (mm) of Cranial Parameters for Cormohipparion skinneri, Gidley Horse Quarry, Clarendon Beds, Medial Clarendonian, Texas Characters illustrated in figures 3–Figure 45.

In ventral aspect, the incisor arcade is smoothly rounded (best seen in F:AM 73920), the incisors having cement-filled infundibula (also in F:AM 73910). The canines are situated about midway between I3 and P2 (fig. 21A). DP1 is absent in this adult skull. Premolars are larger than the molars and tend to have a somewhat more complex enamel pattern, comparable to the general situation in the C. occidentale group. The dentition in F:AM 73912 (fig. 21B) is in a relatively mature state of wear (P2 seems to be more than half worn; table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }), but the bifid versus the single pli caballin distinguishes premolars from molars. At this wear stage, the pre- and postfossettes of P2 are separate rather than confluent along their labial margins; cheek tooth protocones are elongate and isolated but relatively broad and slightly concave lingually in the premolars versus nearly flat in the molars.

The lower dentition and mandible is not represented in the type material of C. fricki.

Diagnosis

Based on the type and referred material, C. fricki (fig. 22) has a much shorter muzzle (no. 1) than C. occidentale and, apparently, C. merriami, but not C. matthewi or C. skinneri, n.sp., and is longer than in C. johnsoni. Cormohipparion fricki has a much longer palatal region than in the Machaerodus and Hans Johnson samples of C. occidentale, C. matthewi and C. johnsoni, and a shorter overall cranial length (no. 6) than C. merriami, C. skinneri, n.sp., C. matthewi and the XMas-Kat (but not Machaerodus or Hans Johnson samples) of C. occidentale. The choanae and palate are somewhat wider (nos. 12–13) in C. fricki than in C. occidentale. C. fricki apparently is much wider across the frontals (no. 18) than any other species and is wider across the glenoid fossae (no. 19) than C. occidentale, C. matthewi, C. johnsoni, and C. merriami but not C. skinneri, n.sp. The snout (no. 25) is relatively much shorter vertically in C. fricki and C. matthewi than in C. occidentale or the other species. The nasal notch is less incised than in C. occidentale (no. 30) in common with all other species, to varying degrees. The DPOF (no. 33) ranges much smaller in C. fricki than in all but C. johnsoni and C. matthewi, and the rear of the DPOF is much farther from the tip of the lacrimal (no. 39) in C. fricki than in all other species (tables 5Table 6, 9, 11Table 11(Continued), 16).

Relative to C. matthewi (fig. 15), the cranium of C. fricki (fig. 22) differs in having a longer palate (no. 2), a somewhat greater cheek tooth length (nos. 7–9), a somewhat wider palate (no. 13), a distinctly greater separation between the IOF and rear of the DPOF (no. 34), and a much greater separation of the anterior tip of the lacrimal and rear of DPOF (no. 39, table 9 versus table 14). C. fricki also differs in having relatively shorter palatal fenestrae (no. 3), a shorter postpalatal region (no. 5), a distinctly wider diastemal region (no. 14), a relatively wider muzzle (no. 15), an apparently wider frontal region (no. 18) and transglenoid width (no. 19), a higher DPOF (no. 35) and IOF-alveolar border distance (no. 37), and a medially deeper DPOF (no. 40). The two species are comparable in snout length (no. 1) and height (no. 25), basicranial length (no. 4), total cranial length (no. 6), choanal width (no. 12), orbital dimensions (nos. 28–29), nasal notches (no. 30), facial lengths (no. 31), POBs (no. 32), and DPOF lengths (no. 33) and facial crest elevation (no. 36), and DPOF-alveolar border distance (no. 38).

Cormohipparion fricki (fig. 22) differs from C. johnsoni (fig. 18) in having a distinctly longer muzzle (no. 1) and palate (no. 2), longer premolar (no. 7) and cheek tooth (no. 9) lengths, longer facial region (no. 31) and POB (no. 32), and a longer IOF–DPOF distance (no. 34), as well as larger orbits (no. 28). The lacrimal is much more widely separated from the rear of the DPOF (no. 39, table 11Table 11(Continued), versus table 14), and the IOF is more widely separated from the alveolar border (no. 37). C. fricki also differs from C. johnsoni in having a wider palate (no. 13), diastema (no. 14), muzzle (no. 15), fontal region (no. 18), and transglenoid width (no. 19). Relative to C. johnsoni, C. fricki has a greater vertical dimension for the DPOF (no. 35), as well as vertical separation of the IOF from the alveolar border (no. 37), in contrast to the DPOF actually lying nearer the alveolar border (no. 38). C. fricki and C. johnsoni are comparable in snout height (no. 25), postpalatal dimensions (no. 3), molar lengths (no. 8), relatively short length of the DPOF (no. 33), depth of nasal notches (no. 30, probably in general context to no. 25), and medial depth of the DPOF (no. 40).

In comparison with C. merriami (fig. 20), C. fricki (fig. 22) differs in having a much shorter muzzle (no. 1) and palate (no. 2) and overall cranium (no. 6), a somewhat shorter facial region (no. 31), a much lower snout height (no. 25), somewhat longer (no. 28) but equally tall (no. 29) orbits, a distinctly longer POB (no. 32), and an apparently greater separation of the lacrimal from the rear of the DPOF (no. 39; table 11Table 11(Continued) versus table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). C. fricki also differs from C. merriami in having a distinctly wider frontal (no. 18) and transglenoid region (no. 19). These two species are comparable in postpalatal length (no. 3), basicranial length (nos. 4–5), cheek tooth lengths (nos. 7–9; slightly shorter in C. fricki), palatal width (no. 13), muzzles (no. 15), likely diastemal width (no. 14), low inion height (no. 22), and small nasal notch incision (no. 30). Based on an overlap in 1 S.D. ranges, the two species apparently are comparable in having a similar DPOF length (no. 33), IOF positions (no. 34), DPOF heights (no. 35), facial crest–DPOF positions (no. 36), IOF–alveolar border distances (no. 37), and IOF–DPOF distances (no. 38). Also, the DPOFs are comparably deep medially (no. 40).

Referred Material

MacAdams Quarry: F:AM 73913, snout and facial region, right and left I1, LI2, RI3, right and left C, LdP1–M3, RP2–M3, subadult male, M3 erupting. F:AM 73914, partial cranium, LI2, right and left dP1–4, M1, juvenile female, M1 erupting. F:AM 73915, partial cranium, RI2–3, right and left C, P2–M3, adult male, M3 medial wear. F:AM. 73920, a partial cranium with right and left I1–2, LI3, right and left C, dP1–M3, subadult male, M3 early wear. F:AM 73921, snout and facial region, right and left I1–3, C, P2–M3, adult female, M3 well worn. F:AM 73925, facial region, RP3–M3, LP2, M3 early wear. F:AM 73938, cranial fragment.

Hollow Horn Bear Quarry, Ash Hollow Formation undifferentiated (Skinner and Johnson, 1984; figs. 2A, 3; table 1), Todd County, South Dakota: F:AM 71880, partial cranium with right and left P2–M3; F:AM 71881, LdP2–4, M1 erupting; F:AM 71873, juvenile cranium and mandibles with right and left dP1–P4, M1 erupting, right and left dp1–4; F:AM 71874, juvenile palate with right and left dP1–4; F:AM 71875, LP2–M3, late wear; FAM: 71876, R maxilla with dP2–4, M1, M2 erupting; F:AM 71879, dorsally flattened skull with RI1, right and left C1, P2–M3.

Description of Referred Material

Based on the material from MacAdams Quarry, major features of the cranium are comparable to those of the type specimen. The upper cheek tooth dentition is relatively complex (fig. 21B). As indicated in table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }, the tallest P2 is about 41 mm tall and P3 is about 46 mm tall in F:AM 73913, which show an incipiently developed occlusal pattern. Based on this specimen, the unworn MSTHT of P2 likely was about 45 mm and that of P3 was likely about 55 mm, from which it may be suggested that P4–M2 also were about 55 mm tall in the unworn condition, with M3 being somewhat shorter. If this is reliable, it suggests that C. fricki was somewhat lower crowned than the XMas-Kat C. occidentale (table 3B). The upper cheek tooth dentition of C. fricki also appears to have been somewhat less complex than in the XMas-Kat C. occidentale, especially as regards the anterior border of the prefossette, which usually bears only 1–2 plis after about medial wear. Similarly, the mean values of plis on the opposing borders of the pre- and postfossettes are 5∶4, 9∶5, 7∶4, 7∶6, and 7∶5 for P2–M2, respectively, based on the relatively small sample (table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). The labial margins of the pre- and postfossette in P2 are separate rather than confluent, except for F:AM 73925. The simple morphology of P2 in C. fricki contrasts with the more complex morphology in the type of C. occidentale. Similarly, the commonly spurred hypoconal groove in P2–4 of C. fricki is more complex than in XMas-Kat and in the type of C. occidentale.

As indicated in F:AM 73913, dP1 persists into early adult wear in which M3 is in the process of being erupted, but P2–3 are in early phases of wear, with MSTHTs of about 41 mm and 46 mm, respectively (table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). Based on F:AM 73913 and F:AM 73920, the ratio of the length of dP1 to the length of P2 is 0.38–0.41 (table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). This compares with a ratio of 0.3 in C. johnsoni (table 12Table 12 (Continued)), 0.37 in C. merriami (table 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] }), and 0.3 in juvenile XMas-Kat specimens of C. occidentale (table 7{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }). DP1 is unknown in adult C. matthewi and C. skinneri, n.sp. The relatively large size of dP1 in C. fricki and C. merriami appears to be plesiomorphic in those species. The dP1 is present in about 60% of the specimens from MacAdams Quarry, regardless of sex. It also is absent in both female and male crania.

Upper cheek tooth protocones tend to be relatively elongate as in the type specimen, except in F:AM 73915, where they are shorter (table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). Premolar plis caballin in C. fricki appear to be about as complex as in C. occidentale but less complex in the molars (tables 7–{ label needed for table-wrap[@id='i0003-0090-306-1-1-t15'] }8), comparable to the other species of the C. occidentale group. Except where not applicable (deciduous premolars; two specimens), the pre- and postfossettes of P2 are about equally confluent or separate in the hypodigm (N  =  3 for each).

F:AM 71880 (fig. 8) is a partial skull from the Hollow Horn Bear Quarry that preserves the anterior portion of the orbits, the facial region containing the DPOF and palate wherein right and left P2–M3 are retained. As indicated in table 3B, the measured crown height of P2 in F:AM 71880 is within the range for C. fricki (table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). The upper cheek tooth morphology of F:AM 71880 is compatible with that for C. fricki (figs. 8B, 21B), considering the earlier wear stage of the Hollow Horn Bear Quarry specimen. The nearly unworn M3 of this specimen is likely to have been about 37 mm tall (table 4B) and about 45 mm in the unworn condition. This suggests that the unworn crown heights of the tallest cheek teeth (P4, M1) would have been on the order of 55 mm, comparable to the MacAdams sample of C. fricki. The unworn MSTHT of M1 and M2 in F:AM 71876 is circa 55 mm and 50 mm, respectively (table 3B). Specimens from the Hollow Horn Bear Quarry show a confluent pre- and postfossette in P2 in three specimens (not applicable in four) and separate in none.

F:AM 71880 preserves the facial part of the cranium (fig. 8A), which displays a well-developed DPOF that is situated well anterior to the orbit and to the anterior tip of the lacrimal bone. Note that the lacrimal is separated from the maxillo-nasal suture by a jugo-nasal suture that is about 5 mm long. The lacrimal therefore does not reach the maxillo-nasal suture. The DPOF is slightly pocketed posteriorly and has a well-developed anterior rim adjacent to, but above, the IOF. The IOF is located above the boundary between P2 and P3. Other specimens from the Hollow Horn Bear Quarry (referred material) demonstrate a comparable morphology.

Discussion

Based on these specimens, the unworn crown height for C. fricki is on the order of 55 mm for M1 and somewhat less for P2 and M3 (table 3B). The anterior border of the prefossette and the opposing borders of the pre- and postfossettes are relatively complex, even in medial wear (e.g., F:AM 73925, F:AM 73901, F:AM 74915, table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }). For the overall sample, premolar plis caballin tend to be at least double; protocones are ovate, commonly lingually concave, and remain isolated from the protoloph in late wear; at comparable wear stages, cheek teeth tend to be relatively wide. P2 pre- and postfossettes are confluent in all three of the specimens in which it can be appraised (not applicable in four).

Webb (1969: 102) allocated material from Little Beaver B and Little Beaver A sites, Minnechaduza Fauna, Nebraska, to Neohipparion occidentale. As revised by Skinner and MacFadden (1977), this species is referable to Cormohipparion. In the present revision, the Minnechaduza material is referred to C. fricki. This is based chiefly on the crown height of the upper cheek teeth (unworn MSTHT 51–56, according to Webb, 1969: 102). This referral also is compatible with the degree of enamel pattern complexity and the relative elongation of the protocone of the Minnechaduza sample as described by Webb (1969). The Minnechaduza sample does not reach the crown height and enamel complexity characteristic of C. occidentale as described herein.

At present, no adult mandibles or lower dentitions can be directly associated with crania referred to C. fricki. Based on the crown height of the upper cheek teeth, the confluence of the pre- and postfossettes of P2, the ovate protocone proportions in the upper cheek teeth, general cranial characters, and the retention of dP1, C. fricki possesses many features likely prophetic of the North American ancestry of the Hippotheriuim Datum. This is discussed further below.

Cormohipparion skinneri, new species

figures 23–24; tables 2–3Table 3 (Continued)Table 3 (Continued), 6Table 6 (Continued), 16–Table 17{ label needed for table-wrap[@id='i0003-0090-306-1-1-t30'] }18

Type Specimen

F:AM. 73909, nearly complete undistorted cranium and mandible, with right and left I1–3, C1, P2–M3, right and left i1, Li2–3, right and left c1, Rp2–m3, Lp3–m3.

Type Locality

Gidley Horse Quarry, Clarendon beds, Texas.

Age

Medial Clarendonian, Texas.

Distribution

Type locality.

Description of Type Specimen

F:AM 73909 is comparable overall to C. occidentale. The type cranium is slightly crushed at the frontal region so that the facial area anterior to the orbit is distorted. The nasal bones and the dorsal border of the facial region are preserved, but the frontal region is somewhat depressed, and the parietals are broken. The lateral surface of the face above the facial crest is depressed so that the lacrimal morphology is somewhat distorted. The dorsal half of the lacrimal is best preserved on the left side. The lower course of this element is partially obscured on the right side, as well. The morphology shown in figure 23A is a composite of both sides. In comparison, F:AM 71800 from the XMas-Kat sample is less distorted. The upper cheek tooth dentition is well preserved, as is the mandible and lower dentition. The angular region of both mandibles is missing (restored), as is the tip of the left coronoid process.

In overall morphology, the cranium of C. skinneri is comparable to that of many contemporaneous equids. The nasal notch extends posterior to the canine, but the relatively elongate facial region as compared with, for example, C. quinni (fig. 26; see also Woodburne, 1996b) is shown by the anterior margin of the orbit being located above the anterior edge of M3 in adult wear. In lateral view (fig. 23A), the overall profile is relatively regular, with the dorsal surface gently recurved. Although depressed postmortem, the frontal bones do not appear to have been domed dorsally. The orbits are relatively large (table 16), and the POB is relatively wide. The DPOF is thus well anterior to the orbit, is higher posteriorly than anteriorly but not as distinctly teardrop-shaped as, for instance, in the XMas-Kat C. occidentale. As in Cormohipparion generally, the IOF is located just below the anteroventral margin of the DPOF, usually slightly posterior to the actual anterior end of the fossa. In C. skinneri, the IOF is situated above the rear half of P2, comparable to the condition in C. fricki. The anterior end of the DPOF is not as pronounced as typically seen in XMas-Kat materials of Cormohipparion. The anteriorly wide POB (no. 32, table 16) is reflected in the lacrimal usually being exposed on the posterior three-fifths of the bar (length 2, table 3A) and not entering the rear of the DPOF, as recorded in C. quinni. As in C. quinni and the XMas-Kat materials, the facial crest ends above the premolar–molar boundary at an elevation approximately midway between the alveolar border and the ventral margin of the DPOF. In spite of the wider POB, the DPOF is still situated about as in C. quinni; the rear of the DPOF still is located essentially above the mesostyle of M1. In C. skinneri, the lacrimal is longer than in C. quinni, but the extension of the POB appears to have taken place as well between the lacrimal tip and the rear of the POB. This is reflected in dimension no. 32 being about twice that in C. quinni (table 16 in comparison with Woodburne, 1996b: table 16 no. 32), even though the facial length of the lacrimal (nos. 32–39, table 16) is 39 mm in C. skinneri versus about 25 mm in C. quinni. In contrast to that in C. fricki, the lacrimal in C. skinneri is blunt anteriorly and nearly vertical. The maxillo-lacrimal and maxillo-jugal sutures thus are effectively aligned with one another. In addition, the lacrimal appears to have been more broadly exposed on the face (length 2, table 3A), and has a blunt, rather than pointed, anterior end. Consequently, the lacrimal does not extend appreciably anterior to the maxillo-jugal suture.

Figure 26

Cormohipparion quinni, AMNH 108535, MacAdams Quarry, Clarendon beds, Texas, medial Clarendonian. A, left lateral view of cranium. B, occlusal view of left P2–M3.

In ventral aspect, the incisor arcade is smoothly rounded, the incisors having cement-filled infundibula. The canines are sited about two-thirds the distance between I3 and P2. DP1 is absent. The premolars are generally larger than the molars and tend to have a slightly more complex enamel pattern (fig. 23B). The dentition of F:AM 73099 is in a mature stage of wear (table 17{ label needed for table-wrap[@id='i0003-0090-306-1-1-t30'] }). With a MSTHT of about 30 mm, P2 seems to be one-quarter worn, if this tooth was about 40 mm tall in the unworn state, comparable to C. occidentale. In P2–3, the anterior border of the prefossette is moderately complex (up to 4 plis) but simpler in the more posterior teeth (table 17{ label needed for table-wrap[@id='i0003-0090-306-1-1-t30'] }). The posterior border of the prefossette displays 5–8 plis in the premolars and 4–6 plis in the molars. For the anterior border of the postfossette, comparable numbers are 2–5 plis in the premolars and 4–6 plis in the molars. All have but a single pli on the posterior border of the postfossette. The bifid versus the single pli caballin distinguishes P2, P3, and M1 from the other cheek teeth. At this wear stage, the pre- and postfossettes of P2 are separated rather than confluent along their labial margins; cheek tooth protocones are elongate and isolated but relatively broad and slightly concave labially in both premolars and molars.

The mandibles are well developed, except for missing their angular regions (fig. 23D). As indicated in table 18, the mandible is relatively deep beneath the molars but becomes shallower anteriorly. The lower dentition is remarkable in having a relatively strong (premolars) to strong (molars) penetration of the isthmus by the ectoflexid (see fig. 6 for terminology). In P3 and P4, the base of the metaconid and metastylid are constricted, respectively, into right and left halves of the upper arms of an “X” pattern, as shown in figure 23C. In M1 and M2, the metaconid and metastylid respectively form the right and left halves on an ‘H’ pattern, with the cross bar of the H formed by the anteroposteriorly elongate labial end of the ectoflexid (fig. 23C). Protostylids are present on p3–m3. The ectoflexid development is exhibited by a specimen judged to be in adult wear (upper dentition about 25% worn) and is present in all molars, as well as in the premolars. If borne out by additional specimens, if and when recovered, this would represent an evolutionary novelty not found in other species of Cormohipparion.

Diagnosis

Cormohipparion skinneri is distinguished from all other species of the subgenus Cormohipparion in having the characteristic strong development of premolar and molar ectoflexids and the development of the “X” and “H” patterns described above (assuming the type is representative of its species population). As shown in figure 24, C. skinneri is represented only by a single specimen in parameters 1–25 and by two specimens in parameters 28–40. Parameters in which C. skinneri falls within the range indicated on figure 24 for C. occidentale from XMas-Kat Quarry and in figure 13 from the XMas-Kat, Machaerodus, and Hans Johnson quarries are not considered further here. In that context, C. skinneri differs from C. occidentale in having a somewhat shorter muzzle (no. 1), a somewhat longer palate (no. 2), and a much shorter postpalatal dimension (no. 3), even though the basicranial dimension (no. 4) is similar. Cormohipparion skinneri also differs from C. occidentale in having longer premolar, molar, and cheek tooth row lengths (nos. 7–9); a wider choanae (no. 12), palate (no. 13), and muzzle (no. 15); and a very much wider transglenoid dimension (no. 19), in spite of the frontal width (no. 18) being comparable. Although the standard deviation for C. skinneri overlaps that for C. occidentale, C. skinneri has a somewhat larger orbit (no. 29), a somewhat longer facial region (no. 31), and a somewhat taller DPOF (no. 35). Cormohipparion skinneri also differs from C. occidentale in having a much less incised nasal notch (no. 30), a shorter DPOF (no. 33), and a much greater separation of the rear of the DPOF from the alveolar region (no. 38).

In comparison with C. matthewi (fig. 15), C. skinneri (fig. 24) has a shorter muzzle (no. 1), a much longer palate (no. 2), a much shorter postpalatal dimension (no. 3), a smaller basicranial dimension (nos. 4–5) (but a relatively longer cranium overall [no. 6]), longer cheek tooth row dimensions (nos. 7–9), a wider choanae (no. 12) and palate (no. 13), a much wider snout (no. 14) and muzzle (no. 15) (but a narrower frontal [no. 18] versus a much wider transglenoid region [no. 19]), a much taller snout (no. 25), comparably sized orbits (nos. 28–29), comparably weakly incised nasal notches (no. 30), a somewhat longer facial region (no. 31), a longer POB (no. 32), a comparably short DPOF (no. 33), a much greater separation between the IOF and rear of the DPOF (no. 34) and a much taller DPOF (no. 35), a greater separation of the facial crest and the DPOF (no. 36) (but a comparable separation of the alveolar border and IOF [no. 37]), a very much taller DPOF relative to the alveolar border (no. 38), a somewhat greater separation of the anterior tip of the lacrimal and rear of the DPOF (no. 39; table 16), and a somewhat medially deeper DPOF (no. 40).

In comparison with C. merriami (fig. 20), C. skinneri (fig. 24) has a somewhat shorter muzzle (no. 1), a comparable palatal length (no. 2); a shorter postpalatal (no. 3) and basicranial region (no. 4); and a comparable dimension no. 5, overall skull length (no. 6), cheek tooth rows (nos. 7–9), and palatal width (no. 13), but a relatively wider snout (no. 14) and muzzle (no. 15). In contrast, the frontals are comparably wide (no. 18), but the transglenoid dimensions are distinctly greater (no. 19) and the inion height smaller (no. 22) in C. skinneri, with the snout height (no. 25) being comparable in the two species. The orbit is somewhat smaller in C. merriami (nos. 28–29); the nasal notch is somewhat more incised (no. 30); and the facial region is longer (no. 31). The two species have a comparably short POB (no. 32) and DPOF (no. 33), whereas the IOF location (no. 34), DPOF height (no. 35), and facial crest–DPOF separation (no. 36, strongly) are greater in C. skinneri. The IOF–alveolar border separation (no. 37) and DPOF–alveolar border separation (no. 38) are comparable in the two species, whereas the DPOF is farther from the lacrimal (no. 39, table 16), but the DPOF is shallower in C. skinneri.

In comparison with C. johnsoni (fig. 18), C. skinneri (fig. 24) has a longer muzzle (no. 1) and palate (no. 2); a shorter postpalatal dimension (no. 3); longer cheek tooth rows (nos. 7–9); a wider palate (no. 13), snout (no. 14), muzzle (no. 15), frontal (no. 18), and transglenoid (no. 19); a taller snout (no. 25); and larger orbits (nos. 28–29), but a comparably little incised nasal notch (no. 30). The facial region (no. 31) and POB (no. 32) are longer in C. skinneri, but the DPOF (no. 33) is comparably short in the two species. The IOF location (no. 34), DPOF height (no. 35), facial crest–DPOF separation (no. 36), IOF–alveolar border separation (no. 37), DPOF–alveolar border separation (no. 38), and separation of DPOF and anterior lacrimal tip (no. 39; much greater; see table 16) are greater, but the medial DPOF depth (no. 40) is smaller, in C. skinneri.

In comparison with C. fricki (fig. 22), C. skinneri (fig. 24) has a similarly short snout (no. 1), similar palatal length (no. 2), shorter postpalatal dimension (no. 3), and somewhat shorter basicranium (nos. 4–5), but a somewhat longer skull overall (no. 6). The cheek tooth rows are longer in C. skinneri (nos. 7–9), but the choanae (no. 12) and palates (no. 13) are about equally wide in the two species. Cormohipparion skinneri has a somewhat wider snout (no. 14) and muzzle (no. 15) but a much narrower frontal (no. 18) and a slightly wider transglenoid dimension (no. 19), as well as a slightly lower inion height (no. 22). Cormohipparion skinneri has a great deal taller snout (no. 25) and a somewhat longer orbit (no. 29), but a comparably weakly incised nasal notch (no. 30). The facial region (no. 31), POB (no. 32), and DPOF (no. 33, relatively short) are about equally long in the two species, but the IOF is relatively more anterior (no. 34), the DPOF is taller (no. 35), the facial crest is farther from the DPOF (no. 36) (but the IOF is nearer the alveolar border [no. 37]), the DPOF is farther from the alveolar border (no. 38), the anterior tip of the lacrimal is much nearer the rear of the DPOF (no. 39, table 16), and the DPOF is medially shallower (no. 40) in C. skinneri.

Referred Material

F:AM 73910, distorted cranium with right and left I1, LI2–3, right and left P2–M3. Gidley Horse Quarry, Clarendon beds, Texas.

Description of Referred Material

F:AM 73910 is badly distorted in having been crushed dorsoventrally and skewed labially. It seems to have the same basic morphology as in the type specimen, including a lacrimal that is narrower anteriorly than at the orbit but that appears to be blunt anteriorly. At this mature stage of wear (P2 appears to be about one-half its unworn MSTHT), the cheek tooth protocones still are elongate and isolated. In P3 and P4, the lingual margin of the protocone is slightly concave; otherwise, this protocone border is slightly convex to flat in the other teeth. Plis caballin are more complex (trifid) in P2 and P3 than in P4–M3 (bifid, table 17{ label needed for table-wrap[@id='i0003-0090-306-1-1-t30'] }). The dP1 is absent in these adult crania.

Discussion

As developed further below, C. skinneri is most similar to C. occidentale, with which it is contemporaneous, but differs in the features indicated in the Diagnosis. The Texan occurrence of C. skinneri relative to the more northern distribution of C. occidentale suggests that C. skinneri may have pertained to a population characteristic of a southern province.

Cormohipparion sp.

Two samples are referred to Cormohipparion sp., unallocated. The first is LACM 150080, from Punchbowl Formation, Devil's Punchbowl, Valyermo, California. The second is derived from the Dove Spring Formation, El Paso Basin, California.

Valyermo

The locality (V, fig. 1) and stratigraphic setting for LACM 150080 are presented in Woodburne (2005). As discussed there, the complexity of the fossette borders and cheek tooth plis caballin, the subovate outline of the protocone (rather than very elongate; fig. 25) and its likely moderate unworn MSTHT (ca. 50–55 mm) are prophetic of the morphology displayed in upper cheek teeth of Hippotherium primigenium. LACM 150080 seems to be morphologically the closest to the Old World taxon among North American samples. Note that the preservation of the specimen apparently masks the complexity of the plis caballin in P4–M2. The connection of the protocone to the protoloph in P2 at about 50% wear may be a relatively plesiomorphic retention, but, as discussed below, it also occurs in other species of Cormohipparion as well as in Old World material referred to Hippotherium sp. and H. primigenium. In the current report, the age of LACM 150080 is suggested as medial Clarendonian, although an early Clarendonian age cannot be ruled out.

Figure 25

Hippotherium and Cormohipparion. A–D, Hippotherium sp., Pannonian C?, Mariathal, Austria. E, Hippotherium sp., Pannonian C, Gaiselberg, Austria. F–G, Cormohipparion sp., Punchbowl Formation, LACM 7502, medial Clarendonian, California. A, PIUW 3504/3, LP2, reversed. B, PIUW 3540/97, LP3, reversed. C, PIUW 3540/54, LP4, reversed. D, PIUW 3540/136, RM1. E, NHMW 0024/26, RM2. F, partial palate with right and left P2–M2, occlusal view. Preservation of the specimen apparently masks the complexity of the plis caballin in P4–M2. G, right lateral view of partial palate, showing remnant of posteroventral border of DPOF and the IOF located above the anterior root of P3.

As indicated in tables 13{ label needed for table-wrap[@id='i0003-0090-306-1-1-t24'] } and 21 and in figure 25, the complexity of the fossette borders (especially at the posterior border of the prefossette and the anterior border of the postfossette in P3 and M1) rivals or exceeds that of C. occidentale as well as samples of Hippotherium primigenium. At the same time, the ovate protocones and their width-to-length ratios are generally comparable to those seen in samples of H. primigenium. The confluence of the pre- and postfossettes in P2 is seen also in those taxa, as well as in other species of Cormohipparion, apparently most abundantly in C. fricki.

Table 21

Dental Statistics of Upper Cheek Teeth of Hippotherium and Cormohipparion Species

El Paso Basin

Whistler and Burbank (1992) described the litho-, magneto-, and biostratigraphy and geochronology of the fossiliferous succession of the Dove Spring Formation of the El Paso Basin, California (E, fig. 1). Whistler and Burbank (1992: fig. 4) recorded C. occidentale as occurring from the base of the Cupidinimus avawatzensis/Paracosoryx furlongi Assemblage Zone to just within the base of the Paronychomys/Osteoborus diabloensis Assemblage Zone. Based on the specimens discussed below, this range is modified to about 11.8–10.3 Ma, as shown in figure 2. All of the specimens are isolated teeth (or, in one case, in a lower mandible) and have no associated cranial material. Thus, for the purposes of this report, the specimens are allocated as Cormohipparion sp. but, as also discussed below, it appears unlikely that they would pertain either to C. occidentale as defined herein or to the Valyermo taxon.

As summarized in table 19, these teeth overall are of larger size, greater dental complexity, and, likely, taller unworn MSTHT than Hipparion forcei and Hipparion tehonense, as documented by MacFadden (1984), which also occur in the Dove Spring Formation (Whistler and Burbank, 1992: fig. 4). The lower dentition, LACM 138755, shows protostylids present in p3–m3, also characteristic of Cormohipparion. Most show elongate protocone proportions, with only LACM 140854 being more nearly ovate than elongate. In general, crown height (in which the unworn MTSHT could have been on the order of 50–55 mm), protocone index, and fossette complexity in these specimens is similar to the XMas-Kat materials of C. matthewi, but further material from each site is needed before a more definite comparison can be made. In that the material spans an interval of nearly 1 m.y., it is possible that more than one taxon could be represented.

Cormohipparion quinni

figure 26; table 20

Table 20

Cranial Measurements (mm) of Cormohipparion quinni, MacAdams Quarry, Clarendon Beds, Medial Clarendonian, Texas

Figure 26 shows a lateral view of AMNH 108535 from MacAdams Quarry, Clarendon beds, Texas. This specimen also was figured as C. sphenodus by MacFadden (1984: fig. 127). Woodburne (1996b) replaced C. sphenodus with C. quinni. The purpose of illustrating this specimen in the current context is to affirm the presence of this species, with its characteristically short POB (no. 32, table 20) and lacrimal eroded by the rear of the DPOF, as well as relatively simpler cheek tooth dentition, in contrast to the C. occidentale group of taxa described here. F:AM 108535 confirms the presence of this species in the MacAdams Quarry sample, and the persistence of this species, as shown in figure 2, is contemporaneous with medial Clarendonian descendants of the speciation event responsible for the genesis of the C. occidentale group of taxa.

Results

The Austrian material, mostly (?or completely) recovered from sites correlated with the Pannonian C Stage, appears to be somewhat lower crowned and morphologically conservative relative to German and Austrian specimens allocated as H. primigenium and, in Spain, as H. koenigswaldi. In that the Eppelsheim sample from the Dinotherium Sands apparently was drawn from the same general site as produced the type specimens of H. primigenium Meyer 1829 (Franzen et al., 2003b), the Austrian material may warrant a different, presumably specific, name. Resolution of this taxonomic situation is beyond the scope of the present work, which focuses on the morphology likely to have characterized the colonizers of the Old World as stemming from Cormohipparion. The rather limited dental material from sites of Pannonian C age affects the resolution of this question to some degree, but until proved otherwise, the working hypothesis of this report is that the Austrian samples from Gaiselberg, Atzelsdorf, and possibly Mariathal show the upper-cheek-tooth morphology of the geologically oldest representatives of the colonizing population of Hippotherium. Further study may reveal the specific allocation of the Inzersdorf and Laarberg crania (tables 22Table 22 (Continued)Table 22 (Continued), 24), if different from Hippotherium sp. as utilized here. If these two crania (table 24) may be considered to represent the same population, it seems to be have a somewhat shorter muzzle length (no. 1), a possibly smaller upper-molar length (no. 8), a shorter orbit (no. 28), a somewhat shorter facial region (no. 31), and a somewhat shorter DPOF (no. 33) compared with the sample from Höwenegg.

As indicated in table 21, it appears that the unworn MSTHT for P4 and M1 in the samples of Pannonian C age (Gaiselberg, Atzelsdorf), of C/E? age (Mariathal) and of D/E age (Inzersdorf; Daxner-Höck, 1996) is somewhat lower crowned than the H. primigenium material from the Dinotherium Sands (Eppelsheim), Germany, and H. koenigswaldi from Nombrevilla, Spain. The Höwenegg sample of H. primigenium (Germany) is possibly more like the Austrian material in this regard. In addition, the Gaiselberg material also shows that P2 is characterized by having pre- and postfossettes that are connected in their labial portions. The Gaiselberg specimens (represented in fig. 25E) also appear to have a somewhat lower overall plication count than the Eppelsheim sample. The Mariathal (Austrian) material also appears to have a relatively low plication count, and the pre- and postfossettes of P2 also are mostly confluent (table 21), similar to the condition in the Gaiselberg sample. In their similarity to the Gaiselberg and Atzelsdorf material, the specimens from Mariathal also may prove to be of Pannonian C age. Unfortunately, only the Mariathal taxon is represented by a relatively large sample size, and whereas it may be unlikely that the portions of the paleopopulations represented by the specimens from Gaiselberg, Atzelsdorf, and Inzersdorf are fortuitously conservative, caution is still appropriate when assessing their specific allocation. With this caveat accepted, these Austrian specimens are here taken to characterize the morphology of the colonizing taxon, regardless of its name, and to use that morphology to evaluate its possible North American source.

Bernor et al. (1988) reviewed material considered to be derived from Gaiselberg and allocated by those authors to Hippotherium primigenium. On this basis, Bernor et al. (1988) suggested that 35% (seven) of twenty Gaiselberg maxillary cheek teeth of H. primigenium have an elongate-oval protocone shape, similar to that in Cormohipparion occidentale, s.l. (essentially the material of C. occidentale from the XMas-Kat sample), with three more (15%) having a transitional shape between the oval and lingually flattened protocone morphologies, and that this morphology signified affinity between the two taxa. The frequency of lingually flat protocones obtained here (about 30%; five teeth) for the Gaiselberg sample is comparable to that cited by Bernor et al. (1988). But rather than signifying phyletic affinity to C. occidentale, it seems, instead, that the occasional presence of lingually flat protocones in the Gaiselberg (or any other sample of Hippotherium) is best considered as endemic to the sample in question and to carry virtually no phyletic information relative to the possible ancestry of this taxon from a North American species of Cormohipparion, and specifically not from C. occidentale as represented by the XMas-Kat quarries sample.

In addition, some of the specimens utilized by Bernor et al. (1988) actually pertain to a locality (Laarberg) of Pannonian E age rather than from Gaiselberg. Examination of the actual specimen, as well as that in figure 7 of Bernor et al. (1988), indicates that examples with flat protocones occur in RP2, LP4, and right and left M2–3 of the Laarberg skull, NHMW 1842.VII.6 ( =  A4229; Bernor et al., 1988: fig. 7). P2 and P4 are in mature wear (about 25% worn), but M2 and M3 are in an earlier wear stage. The occlusal pattern on M3 is incompletely formed and so had not achieved a fully developed occlusal pattern. The pattern is obscure in right and left P3 and RP4. In light of its Pannonian E age, the Laarberg specimen appears to be chronologically similar to samples from Höwenegg and Eppelsheim, and the presence of lingually flat protocones is shown to be a limited feature of the upper-cheek-tooth dentition of those two larger samples. In consisting solely of a cranium, the Laarberg sample is of limited use in assessing pattern frequency.

Furthermore, the sparse presence of a lingually flat protocone in the German and in Austrian materials of nominally medial Vallesian age, as well as in the Austrian material of early Vallesian age, suggests that this presumptively elongate-oval pattern was persistent chronologically and was not “rapidly lost” in early Vallesian populations of H. primigenium, as proposed by Bernor et al (1988: 439).

Based on this survey, one may conclude that a resemblance in the protocone shape of H. primigenium to the morphology found in C. occidentale (effectively the XMas-Kat sample) mostly has an intraspecific, including ontogenetic, rather than phyletic source, and that an ovate protocone is typical of the upper-cheek-tooth H. primigenium dentition as far as the specimens reviewed here are concerned. The ovate protocone shape found in C. fricki combines with other characters to illustrate its potential ancestry (sister taxon) to Old World Hippotherium.

In addition to having protocones that were generally ovate (width-to-length ratios at or about 50% in P2–M3), consideration of the morphology of the Hippotherium sp. samples from Gaiselberg, Atzelsdorf, and Mariathal, Austria, suggests that the colonizing taxon was relatively low crowned (ca. 50–55 unworn MSTHT), the pre- and postfossettes were confluent labially in P2, the upper-cheek teeth had a relatively low plication index, and plis caballin that ranged between two (most molars) and three (most premolars). Also, to judge from the Austrian specimens from Laarberg and Inzersdorf, as well as from Höwenegg, Germany, dP1 was relatively well developed and persisted into the adult condition. Based on a single skull from Inzersdorf (NHMW 1875.VI.I), the DPOF was relatively large, the POB was relatively wide, and the anterior tip of the lacrimal was found about halfway across the POB. This is generally comparable to the condition figured in the Höwenegg sample in Woodburne and Bernor (1980: fig. 3B), and in fig. 3.1.1.2 in Bernor et al. (1997). In SMNK Hö A, the lacrimal occupies somewhat more than half the length of the POB and has a blunt anterior end in sutural contact with the maxillary. The vertical anterior end of the lacrimal is somewhat longer and more vertically oriented than in, for example, C. occidentale (fig. 10B). The lacrimo-nasal suture in H. primigenium appears to be more linear and not as concave dorsally, compared with figure 10B. The ventral border of the lacrimal is similar in the two taxa. The shape and sutural relationships, as well as the extent, of the lacrimal bone in the Austrian Hippotherium sp. remains of interest in that the morphology for H. primigenium described above differs from the more anteriorly attenuated condition displayed by the lacrimal of C. fricki (fig. 8A), which is otherwise similar to the Austrian taxon in general crown height, retention of dP1, and high frequency of P2 confluence of the pre- and postfossettes. In being about 11.5–12 Ma old, C. fricki is in a chronologically appropriate position to have been the source of the Hippotherium Datum of the Old World.

As shown in table 24, the Inzersdorf skull is comparable in measurable dimensions to those from apparently equally young (about medial Vallesian) sites in Austria (Laarberg), but both of these are somewhat smaller in some dimensions compared with the sample from Germany (Höwenegg) and generally compatible with a C. fricki size range.

However, C. fricki tends to have a more elongate protocone (a width-to-length ratio generally less than 50% in P3–M1; table 15{ label needed for table-wrap[@id='i0003-0090-306-1-1-t27'] }), as also is consistent with all other species of Cormohipparion at or above node 8 on figure 30 and in contrast to the “more ovate” ratios in the Gaiselberg sample. It also is important to recognize that, except in C. fricki, dP1 is typically absent in adult members of all taxa higher than node 8 in figure 27 and that the Vallesian samples of Hippotherium typically retain this tooth, as also seen in C. merriami.

Figure 30

Cladogram of species of Cormohipparion, the LACM specimen, and Hippotherium primigeniuim with outgroups of Parahippus leonensis, “Merychippus” primus, M. insignis. Tree length is 124; Consistency Index is 0.742; Retention Index is 0.782; Homoplasy Index is 0.258. Bootstrap values are in boldface. Based on a character distribution analysis via MacClade 4.08. Nodes 1–8 are as in fig. 27.At Node 9, the maxillo-lacrimal and naso-lacrimal sutures form an obtuse angle (14∶0); the posterior border of the premolar prefossette has 8–9 plis (23∶4).C. matthewi is distinguished by having a reduced MSTHT to ca. 42 mm (18∶3), and a lower premolar ectoflexid that penetrates the space between the metaconid and metastylid (31∶0).At Node 10, the molar posterior prefossette increases to 10 or more (24∶5); the premolar postfossette plis increase to 2–3 (27∶1); the molar postfossette plis similarly increase to 2–3 (28∶1); and the molar pli caballin is usually double (30∶2)The LACM specimen of Cormohipparion is distinguished by having reverted to a P2 protocone that attaches to the protoloph within 20–30% wear (20∶2); an increase in premolar posterior prefossette plis to 10 or more (23∶5); and an increase in premolar plis caballin to 4 (29∶4).Node 11 has no distinguishing characteristics but is the point at which C. occidentale and H. primigenium diverge from the LACM specimen of Cormohipparion.H. primigenium is distinguished by the IOF located above the P2–P3 boundary (10∶0); the low position of the IOF relative to the ventral border of the orbit (11∶0); the maxillo-lacrimal and naso-lacrimal sutures forming an acute angle (14∶1); the maxillo-lacrimal and lacrimo-jugal sutures forming an acute angle; the dP1/P2 length ratio retained at ca. 60% (17∶1); and their being 8–9 molar anterior postfossette plis.C. occidentale is distinguished in having an unworn cheek tooth MSTHT 60–66 mm (18∶6); their being 8–9 molar posterior prefossette plis (24∶4); and the orbit being posterior to M3 (38∶2).At Node 12, the IOF is located above P2 (36∶1).C. fricki is distinguished by having a palpable anterior rim of the DPOF (8∶1); molars have 2–3 plis on the anterior border of the prefossette (22∶1); the protocone retains a spur (35∶0); and the orbit is posterior to M3 (38∶2).C. skinneri is distinguished by having the lower premolar ectoflexid making an X-pattern with the metaconid/metastylid (31∶2).

Cormohipparion sp. from the Punchbowl Formation of California possibly is similar in MSTHT (estimated) to C. fricki but is derived (based on the single specimen) in the much more ovate cheek-tooth protocones (width-to-length ratio nearly or greater than 60%) and may be independently derived in this feature compared with the other species at or above node 8, as also may be the case for plication complexity of P3–M1, again taking into account that this is a single specimen. However, the clear linking of the pre- and postfossettes in P2 of the California specimen has a strong resemblance to the Atzelsdorf, Gaiselberg, and Mariathal samples of Hippotherium and distinguishes these samples from the later Vallesian localities of Höwenegg and the majority of specimens from Eppelsheim. Future work may show that the Austrian Pannonian C and E samples pertain to a species distinct from H. primigenium, and, perhaps, that the Cormohipparion sp. from the Punchbowl Formation should be referred to it. Inspection of Cormohipparion samples regarding the morphology of P2 indicates that those that have confluent pre- and postfossettes occur about equally (as opposed to the separated condition) in all species except C. skinneri, although the numbers of specimens are small in all samples except for C. occidentale, C. fricki, and C. matthewi (table 21). It appears that more specimens show the confluent versus the separated condition in C. fricki (all were confluent in the Hollow Horn Bear Quarry material, and 60% were confluent in the MacAdams Quarry sample). On this basis, it may be suggested tentatively that of all North American samples, this species is most like the Vienna Basin Pannonian C material in that regard. Still, the at least partial absence of dP1 (see below) and more elongate protocones of C. fricki tend to argue against close affinity with Hippotherium.

Figure 30 is a cladogram of taxa shown in figure 27, with the addition of the LACM specimen of Cormohipparion sp. (LACM) and H. primigenium. The interrelationships of Cormohipparion species are basically those shown in figure 27. The increased plication complexity and crown height relative to C. matthewi (node 10) help link C. occidentale, H. primigenium, and the LACM specimen relative to node 8 or 9, but this means that the loss of dP1 at node 8 would be reversed for H. primigenium, where dP1 is not only persistent but large, a plesiomorphic trait otherwise typical of taxa at nodes 6 and 7. The samples are too small for C. johnsoni and C. merriami to be definitive in this regard, but in C. occidentale, dP1 is absent in all adult skulls from the XMas-Kat quarries, regardless of sex, and when dP1 is present in samples from the Machaerodus and Hans Johnson quarries, the tooth is present only in female crania. In C. fricki, dP1 is present in about 60% of crania of the adult specimens, regardless of sex, and it is scored as present in figure 30. The absence of dP1 in many specimens of C. fricki suggests that the tooth was in the process of being lost from node 8, as was accomplished in C. skinneri, C. matthewi, and C. occidentale. It is plausible that dP1 was present consistently in C. merriami and C. johnsoni, also given its common presence in C. quinni. An increase in premolar pli caballin complexity (score of three; no. 29, table 2) also separates H. primigenium from C. occidentale. In placing H. primigenium as a sister taxon to C. occidentale, the analysis apparently is influenced by the likely convergence between these two taxa as regards general size, an approach by C. occidentale to the fossette-pattern complexity of H. primigenium, and has downplayed the presence of dP1 in H. primigenium versus its loss in C. occidentale. The phyletic relationships for H. primigenium illustrated in figure 29 appear more realistic.

Figure 31 is a log-ratio diagram of cranial parameters of C. fricki (table 14) relative to H. primigenium as represented by the Höwenegg sample of table 24. Compared with C. fricki—and allowing for a potential range of S.D. for these features for which there is only a single specimen, as based on comparison with other samples of the C. occidentale group—the Höwenegg sample of H. primigenium differs in having a strongly longer muzzle (no. 1), a longer skull overall (no. 6), a higher snout in lateral aspect (no. 25), a longer orbit (no. 28), a more deeply incised nasal notch (no. 30), a longer facial region (no. 31), a longer POB (no. 33), a greater rear DPOF–IOF dimension (no. 34), a greater DPOF height (no. 35), and a greater DPOF–alveolus height (no. 38). The basicranial length (no. 4) is comparable in the two samples (tables 14, 24). Figure 31 also shows that the two taxa have a similar orbital height (no. 29; C. fricki is somewhat smaller), so it appears that these differences do not reflect strict allometry relative to basic cranial or body size. A similar analysis relative to the other species of Cormohipparion suggests that parameters nos. 1, 6, 25, 30 (except for C. occidentale), 31 (except for C. skinneri), 32 (for C. occidentale and C. skinneri), 34, and 35 (possibly except for C. skinneri) also differ with respect to H. primigenium, although not always to the same degree of separation shown for C. fricki. Further, as compared with C. johnsoni, the longer premolar length (no. 7); longer total cheek-tooth length (no. 9); greater diastemal width of the muzzle (no. 14) and nasal notch (no. 25); and the longer POB (no. 32) in H. primigenium possibly reflect an allometric source, given the generally smaller cranial size of C. johnsoni. Still, all changes are not equally proportional, and it appears that the cranium of H. primigenium differs from that of all species of Cormohipparion. This, along with dental features, suggests that the base of the Hippotherium radiation in the Old World differs phyletically from members of Cormohipparion in North America. As summarized above, the dental features seen in the Gaiselberg, Atzelsdorf, and, possibly, Mariathal samples of Hippotherium (as well as the geologically younger samples from Höwenegg and Eppelsheim) are derived in the increased complexity of the enamel pattern of the upper-cheek teeth, including the anterior border of the prefossette and in the number of plis caballin. Future analysis may show whether or not Hippotherium is phyletically distinct from Cormohipparion and whether the two taxa retain a separate nomenclature.

Figure 31

Log-ratio diagram of cranial parameters of Hippotheriium primigenium from Höwenegg, Germany, compared with C. merriami. June and Midway quarries, Nebraska.

It is likely that all or most of the features definitive of the Hippotherium clade would have been developed in the Old World, so their similarity to some attributes of C. occidentale, s.s., must be convergent rather than originally shared. The disparity in phyletic linkage and geologic age of H. primigenium versus C. occidentale (fig. 29) also indicates a convergent source for such similarities. Whereas features seen in the well-preserved cranial and dental material of C. occidentale can provide useful general comparisons with elements of the Hippotherium radiation in the Old World, they are of little direct phyletic significance and should be avoided in future phyletic characterizations relative to Old World equids.

It also appears that Cormohipparion was restricted to North America, including Mexico (Ferrusquia-Villafranca, 1990). Bernor et al. (2003) nominate C. sinapensis from deposits in Turkey that range in age from about 10.5 m.y. to 10.0 m.y. The taxon is based on four specimens from at least three quite separate localities, so very little can be learned about intraspecific variation, regardless of the number of species potentially involved. If it can be considered together, the material indicates a taxon about 13–19% smaller than H. primigenium, based on parameters of the cranium, and this may be compatible with the apparently low unworn cheek tooth MSTHT of about 50 mm. Whereas referral of this species to Cormohipparion cannot be absolutely ruled out, the available information also supports its assignment to Hippotherium (as clearly stated by Bernor et al., 2003). Absent a demonstration of any features phyletically distinctive of Cormohipparion, this is followed here, and C. sinapensis is not recognized as a species of Cormohipparion. This carries the implication that there still is evidence for only a single dispersal event from North America to account for origin of Old World hippotherine horses in the early Vallesian of that region, including deposits of Pannonian C age in Austria.

Further work on the Old World samples discussed here may alter these conclusions, but at present it is reasonable to propose that a taxon with the dentition similar to that of the Punchbowl Formation Cormohipparion sp. and having a cranial morphology (and retention of dP1) seen in C. fricki was the most likely source of the Old World Hippotherium Datum and that it was present in North America at about 11.5–12 Ma (fig. 29). If so, it appears that a taxon with this morphology was prevented from dispersing to the Old World until the major sea level fall at Tb3 transpired at about 11.5 Ma (e.g., Woodburne, 2004b: fig. 8.8). Resolution of this question can be sharpened only by the discovery of associated cranial and dental material of the appropriate age on both sides of the Bering Strait. An important result of the present study is that it is no longer relevant to utilize C. occidentale in devising polarities in cranial and dental character states relative to the phyletic origin of H. primigenium.