Etymology: Oreo (Greek): mountain, in reference to Colorado's Rocky Mountains where the talpid was discovered; talpa (Latin): mole.

Type species: Oreotalpa florissantensis gen. et sp. nov., by monotypy.

Diagnosis.—As for the type species.

Etymology: Florissantensis, named for Florissant Fossil Beds National Monument.

Holotype: FLFO 5813 (UCM 99553), a right dentary fragment with m1–m3.

Type locality: UCM locality 92179, Florissant Fossil Beds National Monument, Florissant, Colorado, USA.

Type horizon: Florissant Formation, latest Eocene (Chadronian).

Diagnosis.—Small talpid with m2>m1>m3. Differs from most derived talpids, but similar to some Oligocene talpids, in having ml with metaconid taller than entoconid. Similar in size and morphology to European Myxomygale (Oligocene—Miocene) and Asian Mongolopala (Oligocene). Differs from Oligocene Myxomygale antiqua and M. vauclusensis in more anteriorlyplaced posterior mental foramen (under middle of m1). Differs from M. antiqua, but as in M. vauclusensis and Miocene M. gracilis, m1<m2. Differs from M. gracilis in more posteriorly-placed mental foramen and cristid obliqua does not join metacristid. Differs from M. vauclusensis and Asian Oligocene Mongoloscapter in absence of metastylid and more labial placement of cristid obliqua on posterior wall of metaconid. Differs from Asian Oligocene Mongolopala in m1<m2 and absence of lingual cingulid below paraconid. Differs from late Oligocene North American Quadrudens wilsoni in its smaller size (∼50%), discontinuous labial cingulid, and a smaller, narrower ml that lacks a rectangular trigonid.

Description.—FLFO 5813 (UCM 99553) is a right dentary fragment with m1–m3 (Fig. 2). The posterior mental foramen is positioned under the middle of m1. The apices of several cusps on m1–m2 are broken, but the cusps on m3 are intact and unworn, and consequently we used this tooth as the basis for discussing relative cusp height differences below. The m2 is slightly longer than ml, which, in turn, is longer than m3, as is characteristic of most talpids (Sánchez-Villagra et al. 2006), although in some Oligocene talpids such as Myxomygale antiqua (Crochet 1974) and Mongolopala tathue (Ziegler et al. 2007), m1>m2. On ml, the trigonid is noticeably narrower than the talonid, while they are closer in width on m2 (Table 1). On m3, the trigonid is wider than the talonid, although they are similar in length. The protoconid is the tallest molar cusp, the paraconid is considerably smaller and lower than the proto- and metaconids, and the hypo- and entoconids are similar in height. The metaconid is taller than the entoconid. A postcristid runs between the hypoconid and entoconid, and the hypoconulid is separate. A long cristid obliqua extends from the hypoconid diagonally to the posterior wall of the trigonid lingual to the midline of the tooth and ascending the wall of the metaconid, although not to its apex. A V-shaped talonid notch occurs on m1–m2, as in most living talpids (Sánchez-Villagra et al. 2006), although the meta- and entocristids are high enough to lingually close off a relatively deep talonid basin. The ml and m2 bear a tiny hypoconulid (= entostylid of Hutchison 1974) that is situated directly posterior to and below the entoconid. The m3 has an unreduced talonid that lacks a hypoconulid. The labial cingulid is not continuous around the hypoconid and protoconid. There is no lingual cingulid. An anterior cingulid is present on m1–m3. Although m1–m2 bear a posterior cingulid, m3 lacks one.

Discussion.—Oreotalpa appears morphologically most similar to Oligocene—Miocene Myxomygale from Europe and lower Oligocene Mongolopala from Asia. Incidentally, Ziegler et al. (2007) noted that Mongolopala is morphologically most similar to Myxomygale. Oligocene Myxomygale antiqua and M. vauclusensis differ from Oreotalpa in having a more posteriorly-placed mental foramen (under posterior root of ml). Additionally, M. vauclusensis differs from Oreotalpa in bearing a metastylid and more lingually-placed cristid obliqua (Crochet 1995). Late Oligocene Mongoloscapter also bears a metastylid (Lopatin 2002a). Early Oligocene Mongolopala differs from Oreotalpa in having m1>m2 and a weak lingual cingulid on the paraconid (Ziegler et al. 2007).

Up until discovery of Oreotalpa, the earliest documented North American talpid was the late Oligocene (Arikareean) Quadrodens wilsoni from the Sharps Formation, South Dakota, which was originally described as an erinaceid (Macdonald 1970). Q. wilsoni is about 50% larger than Oreotalpa and bears a relatively larger ml that is wider than m2, with a broad, flattened, rectangular trigonid that comprises almost two-thirds of the tooth. Lower molars of Q. wilsoni have continuous labial cingulids (Macdonald 1970), unlike Oreotalpa, and appear to be more inflated, perhaps adapted to a more durophagous diet.

Because talpid molars are morphologically similar to those of soricids (shrews), chiropterans (bats), and proscalopids, a comparison of Oreotalpa to representatives of these groups is given below.

Soricidae was traditionally considered the sister group to talpids (e.g., Butler 1988). Domnina, the dominant North American soricid during Chadronian time, bears a mental foramen beneath ml and molars that look superficially quite similar to Oreotalpa. However, soricids are more primitive than talpids in having m1>m2>m3 (Sánchez-Villagra et al. 2006). Soricid molars are differentiated from Oreotalpa by having an entoconid that is noticeably taller than the hypoconid and a shallow hypoflexid (Tejedor et al. 2005). Additionally, in soricids, the molar trigonid is longer due to an anteriorly-expanded paracristid and paraconid. Molars bear continuous lingual and labial cingulids, and m3 has a reduced talonid (Repenning 1967; personal observations).

Chiropterans have dilambdodont molars similar to those of several insectivorans, including talpids. However, chiropteran molars are characterized by presence of a complete, continuous labial cingulid (Hand et al. 1994), as well as a hypoconid that is taller than the entoconid, straight entocristid, and a cristid obliqua that terminates on the trigonid wall slightly labial to the midline (Tejedor et al. 2005). Unlike talpids, the posterior mental foramen is more anteriorly placed and under the premolars in bats.

Proscalopids, appearing in the Chadronian with talpid-like teeth, traditionally were placed within Talpidae, although differences in the functional morphology of the skull and forelimb support placement in a separate family (Barnosky 1981). Lower molars of early proscalopids, typified by those of early Oligocene (Orellan) Oligoscalops and Oligocene—Miocene (Whitneyan—Arikareean) Proscalops, bear W-shaped cristids and relatively narrow talonids, especially on m3 where the talonid is considerably narrower than the trigonid (Reed 1961; Barnosky 1981). In contrast to Oreotalpa, Oligoscalops molars lack a distinct anterior cingulid. Lower molars on the holotype of Oligoscalops whitmanensis (CMNH P25800) are worn and incomplete, and their lengths could not be measured. However, in Proscalops, m1 is longer than m2 (Reed 1961), in contrast to talpids where the opposite occurs.

Discovery of Oreotalpa implies that the talpid lineage extends back at least to late Eocene time in North America, which is consistent with some molecular estimates of divergence times that suggest eulipotyphlan subfamilies originated during the Eocene (Douady and Douzery 2003). Phylogenetic analysis based primarily upon nuclear DNA (Roca et al. 2004) suggests that soricids (shrews) and erinaceids (hedgehogs) form a clade that is the sister group to talpids. If correct, this implies that a talpid ghost lineage extends back at least into the Paleocene, based upon the earliest erinaceid, Litolestes, from North America (McKenna and Bell 1997).

From a biogeographic standpoint, appearance of three morphologically similar talpids, Oreotalpa (latest Eocene, North America), Myxomygale (early Oligocene, Europe), and Mongolopala (early Oligocene, Asia) suggests Holarctic intercontinental dispersal of talpids prior to latest Eocene and raises the question of talpid origins. To date, most workers have favored a Eurasian origin primarily because isolated teeth of the earliest Talpidae were found in the Upper Eocene of Europe and assigned to Eotalpa and Geotrypus (Sigé et al. 1977; McKenna and Bell 1997; Whidden 2000), and an undetermined genus and species of Talpinae is known from the latest Eocene of eastern Kazakhstan (Gabunia and Gabunia 1987). Within the biostratigraphic resolution, all appear roughly coeval. Although Oreotalpa does not preclude a Eurasian origin for the Talpidae, it raises the possibility of a North American origin for the group. Proscalopidae, which first appeared in the latest Eocene of North America, has been tentatively considered the sister group to Talpidae by some workers (e.g., Reed 1961; McKenna and Bell 1997). In addition, soricids and erinaceids may form a clade that is the sister group to talpids (Roca et al. 2004). In any event, presence of talpids in the late Eocene of North America, when considered alongside the assumption that sister taxa originate in the same geographic area, provides support for a North American origin for talpids given the antiquity of proscalopids and erinaceids in North America. Biogeography of the oldest soricids is more uncertain. While Domnina from the middle Eocene (Uintan) of North America was long considered the oldest, basal soricid, recent discovery of an Asian middle Eocene soricid may predate Domnina (Lopatin 2002b). The age of these early soricids ultimately rests on refinement of biostratigraphic correlations between Asia and North America.

The Eocene talpid record is currently based upon Oreotalpa from North America (this report), the European genera Eotalpa and Geotrypus (Sigé et al. 1977; McKenna and Bell 1997; Whidden 2000), and an undetermined genus and species from eastern Kazakhstan (Gabunia and Gabunia 1987). Eotalpa is known from two isolated upper molars (RM1 and M2; Sigé et al. 1977) and the holotype of Geotrypus includes a partial mandible with premolars (Hugueney 1972), precluding their direct comparison with the lower molars of Oreotalpa and making it challenging at best to include these taxa in recent talpid phylogenies based mostly on cranial and postcranial characters (Sánchez-Villagra et al. 2006), musculature (Whidden 2000), and molecular sequences (e.g., Shinohara et al. 2003). Our understanding of the geographic origins of talpids requires discovery of many more (and complete) fossils combined with comprehensive phylogenetic analyses that include both fossil and living talpid taxa. Nevertheless, discovery of Oreotalpa indicates that talpids had dispersed between North America and Eurasia by late Eocene time.