Type Species

Gomphos elkema Shevyreva, 1975.

Included Species

Gomphos ellae and G. shevyrevae, new species.

Revised Diagnosis

Similar to other mimotonids but differing from other Glires in having two pairs of lower incisors. Differs from Mimotona (Li, 1977) in its larger size, less transversely extended and more unilaterally hypsodont upper teeth, stronger lophs, a mesostyle usually present, lower molars with a longer trigonid, and a mesoconid on lower molars. Differs from Anatolmylus (Averianov, 1994; Averianov and Godinot, 1998) in having a shallow horizontal ramus. Differs from Mimolagus (Bolin, 1951) in its smaller size and higher-crowned cheek teeth with distinctive cusps and ridges.

Localities and Age

Gomphos elkema is known from Bumbanian faunas in Tsagan-Khushu, Nemegt Basin, and Ulan-Nur Basin of Mongolia (Dashzeveg and Russell, 1988; Dashzeveg, 1988); Huheboerhe, Wulanboerhe, and Nuhetingboerhe (Meng et al., 2004), and Bayan Ulan (Meng et al., 2005) in the Erlian Basin of Nei-Mongol (Inner Mongolia), China. G. ellae is from Tsagaan Khutel, Mongolia (Kraatz et al., 2009). G. shevyrevae is from the Irdin Manha Formation at Huheboerhe, Erlian. Early to Middle Eocene.

Gomphos shevyrevae, new species

Holotype

A right M1 (IVPP V14669).

Paratype

A right m1 (IVPP V14670).

Included Specimens

A right P4 (or P3) (V14671.1), a right M1 (V14671.2), a right M2 (V14671.3), a left M2 (V14671.4), a right m1 (V14672.1), a right m2 (V14672.2), a left m3 (V14672.3), a left calcaneus (V14673), and a left astragalus (V14674).

Etymology

The species name is in honor of N.S. Shevyreva who named Gomphos elkema.

Diagnosis

Differs from Gomphos elkema and G. ellae in having more robust teeth with higher crowns, inflated cusps; upper molars with more posteriorly expanded hypocone and shelf; absence of the ridge connecting the lingual and labial cusps on P4; lower molars with relatively long trigonid and short talonid, reduced mesoconid and hypoconulid, and absence of the mesostylid; further differs from G. elkema in having extra facets on the calcaneus for articulation with the astragalus and navicular (unknown for G. ellae).

Type Locality and Age

Huheboerhe escarpment, Erlian Basin of Nei-Mongol, Middle Eocene lower beds of the Irdin Manha Formation.

Comparative Description

Measurements of all teeth are in table 1. The tooth we identified as a P4 could also be a P3 (fig. 2). However, the anterior border of P3 is usually narrower than the posterior border in Gomphos. This tooth has its anterior portion as wide as the posterior and bears a tiny posterolingual cusp, suggesting that it is likely a right P4. The P4 is a recently erupted tooth that bears no wear. As is typical of Gomphos, the P4 is unilaterally hypsodont and consists of a lingual main cusp, presumably the protocone, and a labial cusp. The lingual cusp is crescent shaped, and its two lophs form the anterior and posterior edges of the tooth. The labial cusp is conical, having a steeper labial surface and a more rounded lingual surface. Unlike P4 of Gomphos elkema, in which P4 is known (Meng et al., 2004; Asher et al., 2005), the ridge connecting the two cusps is absent in G. shevyrevae; instead, there is a weak ridge extending from the labial cusp anterolingually to join the midpoint of the anterior loph. The absence of a connection between the lingual and labial cusps of P4 recalls the condition found in Mimotona (Li, 1977; Li and Ting, 1993).

Figure 2

Crown views of upper cheek teeth of Gomphos shevyrevae from the lower Irdin Manha Formation, Huheboerhe at Erlian, Nei Mongol. A, right P4 (or P3) (V14671.1); B, right M1 (V14669, holotype); C, right M1 (V14671.2); D, right M2 (V14671.3); E, left M2 (V14671.4). Scale bars are 200 µm.

Table 1

Tooth measurements of Gomphos shevyrevae (in mm)

Upper molars have one major lingual and two minor labial roots. The enamel is thick. M1 is so assigned because its hypocone and shelf are less expanded posteriorly than those of M2, and M1 is usually proportionally shorter than M2 (fig. 2). M1 of G. shevyrevae is similar to that of G. elkema in general morphology. However, the M1 and other upper cheek teeth of G. shevyrevae are more robust with more inflated cusps and stronger lophs. The protocone and its anterior and posterior lophs form a V-shaped structure with a slightly rounded apex (the lingual side of the protocone). The angle created by the lophs is about 45°. The lingual portion of the anterior loph may be homologized with the preprotocrista and the labial portion, which is much thinner, with the preparaconule crista. A paraconule (protoconule) is not distinct, but is indicated by a slight inflation where the postparaconule crista joins the anterior loph. Absence of the paraconule is common in basal Glires, both in simplicidentates, such as Matutinia and Rhombomylus (Ting et al., 2002; Meng et al., 2003), and duplicidentates, such as Mimotona and Gomphos (Li, 1977; Li and Ting, 1993; Meng et al., 2004), but it is usually present in taxa that are more closely related to rodents and in basal rodents, such as Tribosphenomys (Meng et al., 1994, 2007b; Meng and Wyss, 2001), Neimengomys (Meng et al., 2007b), Archetypomys (Meng et al., 2007a), and Cocomys (Li et al., 1989).

There is a weak ridge between the paracone and the anterior loph, which is probably the postparaconule crista. The metacone is smaller and more lingually positioned than the paracone. It is connected with the protocone by the posterior loph of the protocone. The latter loph is strong, almost as wide as the metacone. A mesostyle occurs at the posterior base of the paracone, which blocks the labial exit of the narrow trigon basin. The metaconule is present, but is difficult to distinguish in heavily worn specimens. In lightly worn specimens (fig. 2E) a broader wear facet indicates the metaconule that must be higher and broader than the crista in order for such a wear pattern to be created. In G. elkema the posterior loph of the protocone is relatively slimmer and there is usually a restriction labial to the protocone.

The hypocone and the postcingulum are well developed, and both structures are expanded posteriorly. The hypocone is separated from the protocone by a transverse groove, and a similar condition occurs on M2. This condition appears to differ from the morphology prevailing in G. elkema and G. ellae. In well-worn specimens (fig. 2C), however, the protocone and hypocone become confluent. On the lingual surface of the tooth, a vertical groove between the protocone and hypocone exists near the occlusal surface of the tooth, and this feature also seems more pronounced than on upper molars of G. elkema and G. ellae.

M2 is similar to M1 in general morphology, but it is proportionally longer. The mesostyle is reduced, either being incorporated into the posterior base of the paracone or entirely absent. No M3 was collected.

The lower molars have an anterior and a posterior root. The tooth crown is somewhat higher than that of G. elkema and G. ellae. As in the upper molars, the enamel appears thicker than that of G. elkema and G. ellae. The m1 is identified as such because it has a transversely oriented, ridgelike hypoconulid (fig 3). In Gomphos, the lower molar hypoconulid becomes more pronounced posteriorly. The m1 and m2 are very similar except that m2 has a more posteriorly projecting hypoconulid. The lower molars are proportionally wider, and therefore relatively shorter, than those of other species of the genus. Lower molar trigonids are proportionally longer in relation to their corresponding talonid than is the case in other species of the genus. As is typical of Gomphos, the protoconid and metaconid are transversely aligned, and the trigonid has nearly vertical anterior and posterior walls. The protoconid and the metaconid are subequal in size. The protoconid is worn more extensively and is therefore lower than the metaconid. The protoconid sends the paralophid (paracristid) to the anterolingual base of the metaconid and the protolophid (protocristid) to the posterior base of the metaconid. A crescentic trigonid basin is enclosed by the trigonid cusps and lophids.

Figure 3

Lower molars of Gomphos shevyrevae from the lower Irdin Manha Formation at Huheboerhe, Erlian, Nei Mongol. A1–2, crown and lingual views of right m1 (V14670, paratype); B, right m1 (V14672.1); C1–2, crown and labial views of right m2 (V14672.2); D, left m3 (V14672.3). Scale bars are 200 µm.

Lower molar talonids are as wide as the trigonid and are longer than their corresponding trigonid. The hypoconid is as large as the protoconid and is very close to the posterior wall of the trigonid, so that the hypoflexid becomes a narrow groove. The cristid obliqua (ectolophid) is short and the mesoconid appears very weak relative to other mimotonids. The entoconid is low. Unlike the condition in G. elkema and G. ellae, in which a mesostylid is present (Meng et al., 2004; Kraatz et al., 2009), the mesostylid is absent in the new species. Because of the reduction of the mesoconid and cristid obliqua, the talonid basin is a broad, featureless concavity. The hypoconulid is a transverse ridge at the rear of the tooth, being connected with the entoconid by a strong postcristid.

The m3 is little worn and displays several additional minor cuspules on the occlusal surface; these structures would be quickly removed with wear (fig. 3). The m3 differs from m1 and m2 in having a longer talonid and a more pronounced hypoconulid. However, compared to those of G. elkema and G. ellae, the talonid of G. shevyrevae is relatively shorter and the hypoconulid is reduced; it does not form a projection at the posterior end of the tooth.

The astragalus measures 11.34 mm long and 8.82 mm wide (maximum dimensions) and is partly broken at its head (fig. 4A). The morphology of the preserved portion is identical to that of Gomphos elkema (Meng et al., 2004). The bone is ventrodorsally (or anteroposteriorly) flat. The trochlea is transversely broad and shallow. In G. elkema, the lateral and medial rims of the astragalar trochlea appear to be more asymmetrical, with the lateral much larger than the medial (Meng et al., 2004).

Figure 4

Tarsal bones of Gomphos shevyrevae from the lower Irdin Manha Formation at Huheboerhe, Erlian, Nei Mongol. A, dorsal view of left astragalus (V14673); B, dorsal view of left calcaneus (V14674). Scale bar is 1 mm.

The calcaneus measures 20.8 mm long and 8.93 mm wide (maximum dimensions; fig. 4B). It is typical of Gomphos in that the calcaneoastragalar and sustentacular facets are aligned at the same level, similar to that of lagomorphs (Bleefeld and McKenna, 1985), but the bone lacks a calcaneal canal that is characteristic of lagomorphs (Bleefeld and Bock, 2002). The calcaneoastragalar facet is a narrow, long and convex surface in a proximodistal orientation, nearly parallel to the long axis of the bone, whereas the sustentacular facet is rounded and concave. Although the general morphology of the calcaneus is similar to that of G. elkema, it differs from the latter in having a facet on the distal side of the bone, which probably articulates with the navicular and astragalus. This feature does not occur in known astragali of G. elkema (Meng et al., 2004, 2005; Asher et al., 2005), and is most likely a derived feature of G. shevyrevae. However, in Mimolagus, a similar facet is present on the calcaneus, which is in articulation with the astragalus (Bohlin, 1951).

Taxonomy

The specimens described here clearly represent a new species of Gomphos and the youngest species of the genus, and perhaps, the youngest representative of Mimotonidae. It differs from G. elkema and G. ellae in having more robust teeth that are higher crowned with more inflated cusps, upper molars with more posteriorly expanded hypocones and posterolophs, and lower molars that are proportionally shorter, lack mesostylids, and have reduced hypoconulids. In addition, an extra articular facet is developed on the calcaneus.

G. shevyrevae is more similar to G. elkema than to G. ellae in general morphology. The major difference from G. elkema is inflation of cusps and lophs in G. shevyrevae. G. ellae has shorter, therefore transversely longer upper molars, which may be partly attributable to the advanced wear of the known specimen. The M1 (fig. 2B, C) of G. shevyrevae also shows significant wear, but does not display the tendency of transverse elongation present in upper molars of G. ellae. Although worn, which suggests an adult individual, the tooth enamel layer of G. ellae remains thinner than that of G. shevyrevae. Thickened tooth enamel confers a greater resistance to wear, suggesting an ability to process tougher food.

G. shevyrevae indicates that Gomphos, the sister taxon of lagomorphs of modern aspect, was more diverse and has a longer stratigraphic range than was previously thought. The earliest lagomorphs of modern aspect are known from the Early Eocene (Li et al., 2007; Rose et al., 2008; Lopatin and Averianov, 2006), as are the earliest representatives of Gomphos. The fossil record indicates that the divergence between the Gomphos and modern lagomorph lineages was no younger than the Early Eocene. In addition, G. shevyrevae further suggests that a branch of mimotonids coexisted with the earliest lagomorphs at least up to the Middle Eocene. This is further supported by the description of both mimotonids and lagomorphs from the Andarak 2 locality in Kyrgyzstan (Averianov, 1994). The Gomphos lineage may have extended into still younger age if Mimolagus proves to be more closely related to Gomphos than to any other Glires.

Biostratigraphic Implication

Two Asian Land Mammal ages, the Arshantan and Irdinmanhan, have been widely accepted, but their definition and recognition have not always been clear (Meng et al., 2007c). Dashzeveg and Hooker (1997: 136) suggested that the Arshantan “predates the Irdinmanhan and belongs to the earliest Middle or latest Early Eocene.” Lucas and Emry (2001) considered the Irdinmanhan and Arshantan ages as collectively correlative with the Bridgerian, although Luterbacher et al. (2004) correlated the Arshantan with the entire Bridgerian and the Irdinmanhan with the Uintan. Lucas (2001) also regarded the Arshantan land mammal age to be older than the North American middle Bridgerian.

One example of the difficulty in distinguishing these two ages is found in the fauna from the Kholboldchi Formation of Mongolia. Localities that supposedly expose the same formation have produced different fossils, and most of the fossils were typical of both Arshantan and Irdinmanhan faunas. These localities have been considered either Irdinmanhan or Arshantan (Russell and Zhai, 1987; Dashzeveg and Hooker, 1997; Lucas, 2001; Meng et al., 2007c; Kraatz et al., 2009). Similarly, Averianov and Godinot (2005) considered the Andarak fauna to be Irdinmanhan in age, following the definition of the Arshantan and Irdinmanhan (Lucas, 2001). However, because the Andarak mammals are found within a marine section that yields a selachian fauna, Averianov and Godinot (1998, 2005) argued that the Andarak fauna, and thus the Irdinmanhan, be correlated with the latest Early Eocene (late Ypresian).

Part of the current confusion may have stemmed from the original collections of Arshantan and Irdinmanhan faunas, which appear to have been mixed assemblages of fossils from different ages, as pointed out by Meng et al. (2007c). With new stratigraphic and faunal data, clarification of both the Arshantan and Irdinmanhan faunas can be expected. Here, the stratigraphic relationships of Gomphos elkema, G. shevyrevae, and many other associated fossils, are clear (fig. 1). Between the occurrence of Gomphos elkema and G. shevyrevae lies an interval of strata that represents a considerable amount of time and faunal evolution. It is clear that G. shevyrevae is dentally more derived than both G. elkema and G. ellae. This is consistent with the stratigraphic occurrences of G. elkema and G. shevyrevae in the Huheboerhe section. Morphological polarities in the dentition also suggest that G. ellae is older than G. shevyrevae, supporting the view that the Kholboldchi Formation may contain fossils of Arshantan age (Dashzeveg and Hooker, 1997; Meng et al., 2007c; Kraatz et al., 2009).

Mimolagus rodens is an enigmatic member of the Glires that is thought to be from the Early Oligocene of China (Bohlin, 1951). Meng et al. (2004) recognized the similarity between calcanei and astragali of Gomphos and those purported to pertain to Mimolagus, noting that all were more lagomorphlike than rodentlike. A source of uncertainty regarding the possible mimotonid relationships of Mimolagus pertains to its age, because all other mimotonids have been recovered from Middle Eocene or older beds. However, with a revised placement of the Eocene-Oligocene boundary in Asian terrestrial strata (Meng and McKenna, 1998), it seems likely that Mimolagus may be older than previously thought. Considered in light of the discovery of the Irdinmanhan Gomphos at Huheboerhe described here, the apparent temporal gap between Mimolagus and other mimotonids may be significantly less, if not entirely absent.