New fossils, systematics, and biogeography of the oldest known crown primate Teilhardina from the earliest Eocene of Asia, Europe, and North America

Abstract

Omomyiform primates are among the most basal fossil haplorhines, with the oldest classified in the genus Teilhardina and known contemporaneously from Asia, Europe, and North America during the Paleocene–Eocene Thermal Maximum (PETM) ∼56 mya. Characterization of morphology in this genus has been limited by small sample sizes and fragmentary fossils. A new dental sample (n = 163) of the North American species Teilhardina brandti from PETM strata of the Bighorn Basin, Wyoming, documents previously unknown morphology and variation, prompting the need for a systematic revision of the genus. The P₄ of T. brandti expresses a range of variation that encompasses that of the recently named, slightly younger North American species ‘Teilhardina gingerichi,’ which is here synonymized with T. brandti. A new partial dentary preserving the alveoli for P_1-2 demonstrates that T. brandti variably expresses an unreduced, centrally-located P₁, and in this regard is similar to that of T. asiatica from China. This observation, coupled with further documentation of variability in P₁ alveolar size, position, and presence in the European type species T. belgica, indicates that the original diagnosis of T. asiatica is insufficient at distinguishing this species from either T. belgica or T. brandti. Likewise, the basal omomyiform ‘Archicebus achilles’ requires revision to be distinguished from Teilhardina. Results from a phylogenetic analysis of 1890 characters scored for omomyiforms, adapiforms, and other euarchontan mammals produces a novel clade including T. magnoliana, T. brandti, T. asiatica, and T. belgica to the exclusion of two species previously referred to Teilhardina, which are here classified in a new genus (Bownomomys americanus and Bownomomys crassidens). While hypotheses of relationships and inferred biogeographic patterns among species of Teilhardina could change with the discovery of more complete fossils, the results of these analyses indicate a similar probability that the genus originated in either Asia or North America.

Introduction

Crown Primates—the clade including the last common ancestor of all extant primates and all of its descendants—is typically supported with adapiform and omomyiform fossil primates as its most primitive members. The evolutionary relationship between omomyiforms and extant primates has been debated, but they are usually reconstructed as stem tarsiiforms based on the large size of their orbits, foreshortened skulls, anterior foramina magna (consistent with a habitually vertical head posture), elongate tarsals, and small body size, all of which resemble modern tarsiers (Gingerich, 1981, Beard et al., 1991, Dagosto et al., 1999, Rose, 2006, Gunnell et al., 2008). Phylogenetic hypotheses of euarchontan mammals (euprimates, plesiadapiforms, colugos, and treeshrews) often support omomyiforms as most closely related to extant tarsiers and the sister group of anthropoids to the exclusion of adapiforms and strepsirrhines, thereby indicating they are basal haplorhines (e.g., Seiffert et al., 2009, Ni et al., 2016). The oldest omomyiform genus to appear in the fossil record is Teilhardina (Ni et al., 2004, Rose, 2006, Smith et al., 2006, Gunnell et al., 2008, Beard, 2008), and it is the only primate genus found on all three Holarctic continents prior to Homo.

The oldest crown primate fossils (with the possible exception of the poorly known taxon Altiatlasius from Morocco) correspond conspicuously with the Paleocene–Eocene Thermal Maximum (PETM): a brief interval of increased global temperature by ∼5–8 °C correlated with a massive injection of biogenic CO₂ into the atmosphere 56 mya at the Paleocene–Eocene boundary (Kennett and Stott, 1991, Zachos et al., 2001, Zachos et al., 2008, McInerney and Wing, 2011, DeConto et al., 2012, Gradstein et al., 2012). The shift in atmospheric CO₂ concentration marking the PETM is recorded globally by a carbon isotope excursion (CIE) recovered in marine (Zachos et al., 2001) and terrestrial (Koch et al., 1992, Bowen et al., 2001) sediments, leaf-wax lipids (Smith et al., 2007), and mammal tooth enamel (Koch et al., 1992, Secord et al., 2012). The North American PETM fossil record includes a single omomyiform genus—Teilhardina (Gingerich, 1993, Rose et al., 2011)—as well as a single adapiform genus—Cantius (Gingerich, 1986, Gingerich, 1989, Rose et al., 2012). Both genera can be found in earliest Eocene strata in the Bighorn Basin, Wyoming, where Teilhardina appears lower in stratigraphic section, closer to the beginning of the PETM (Rose et al., 2012).

To date, three species of Teilhardina have been recovered globally from strata in which the CIE, indicative of the PETM, has also been documented: T. asiatica from the Hengyang Basin of China (Ni et al., 2004), T. brandti from both the northern and southern Bighorn Basin, Wyoming (Gingerich, 1993, Rose et al., 2011, Rose et al., 2012), and ‘T. gingerichi’ from the Sand Creek Divide section of the southern Bighorn Basin (Rose et al., 2012). Another species, T. belgica, is known primarily from the fluvial Dormaal Sand Member in Belgium (Smith et al., 2006). This locality has received only preliminary study in terms of its carbon isotope composition, indicating that the CIE is present in at least the upper portions of the vertebrate fossil deposit (Grimes et al., 2006). Dormaal has further been argued to correspond to the CIE recorded in the Doel borehole, located ∼85 km to the northwest (Steurbaut et al., 2003, Smith et al., 2006), and to share a remarkably similar fauna to that of the PETM of North America (Smith and Smith, 1996, Smith and Smith, 2001, Smith and Smith, 2010, Smith et al., 1996, Smith et al., 2002, Zack, 2011, Solé and Smith, 2013, Solé et al., 2013). However, biostratigraphy and faunal turnover throughout England and the East Paris Basin have been used to argue that Dormaal and other T. belgica-bearing sites predate the CIE, implying a latest Paleocene age for this taxon with greater antiquity compared to other species of Teilhardina (Hooker, 2015).

A fifth species, T. magnoliana, is known from the T4 sand, which forms the upper component to the Tuscahoma Formation at the Red Hot Truck Stop (RHTS) locality in Mississippi (Beard, 2008). The T4 sand was argued to have been deposited during the PETM on the basis of faunal similarities to the Bighorn Basin and the high abundance of the dinoflagellate Apectodinium (Beard, 2008, Beard and Dawson, 2009), which becomes much more plentiful in association with alterations to sea temperature and the marine carbon cycle during the PETM (Crouch et al., 2001, Crouch et al., 2003, Zachos et al., 2005, Sluijs et al., 2005, Sluijs et al., 2006). Beard and Dawson (2009) argued that the small size of the macroscelidean Haplomylus (Zack et al., 2005, Penkrot et al., 2008) and the presence of the rodent Tuscahomys at RHTS indicate an affiliation of the locality's fauna with that of the brief, early PETM Wasatchian-M faunal zone in the Bighorn Basin (Gingerich and Smith, 2006, Yans et al., 2006). However, the age of the T4 sand has been disputed: the size of Haplomylus from RHTS most closely resembles post-PETM specimens of Haplomylus from the Bighorn Basin (Gingerich, 2010), while Tuscahomys has been found at many intervals during the PETM and early Eocene in Wyoming (Anemone et al., 2012, Rose et al., 2012, Strait et al., 2016). A core located ∼10 km from the RHTS records part of the body of the PETM in angular glauconitic sands, interpreted to correspond with glauconite underlying the T4 sand at RHTS (Sluijs et al., 2014). Given the slow-forming nature of glauconite (Prothero and Schwab, 2004), the interpretation of Beard (2008) or Beard and Dawson (2009) that fossils of T. magnoliana are from the earliest parts of the PETM cannot be correct. Evidence for the T4 sand occurring during the PETM is supplied by the abundance of Apectodinium at the site; however, this dinoflagellate has been documented in post-PETM sediments lacking a negative carbon isotope excursion due to reworking of PETM sediments (Abdelmalak et al., 2016). The location of the RHTS at the paleo-Gulf Coast margin improves the chances of allochthonous sediment mixing during fluvial depositional events. Indeed, Sluijs et al. (2014: figure 3) found elevated abundances of Apectodinium in the lower Hatchetigbee Formation—above the level of the T4 sand or the CIE—disputing that its presence at the RHTS must be indicative of a PETM age. The only certainty conveyed by these data is that the age of T. magnoliana and its associated fauna must occur sometime after the early part of the PETM, but from an unknown time either during or following the climate event in the early Eocene. The presence of Teilhardina and closely related species of other mammalian orders on all Holarctic continents during the PETM is consistent with the hypothesis that PETM climate change facilitated rapid mammalian intercontinental dispersal during this brief window of time (Bowen et al., 2002, Smith et al., 2006, Beard, 2008, Solé and Smith, 2013).

The oldest euprimate species for which dates can be reliably estimated in North America is T. brandti, which historically has been known from few fossils. The species was originally described by Gingerich (1993) based on an isolated right M₂ from the Clarks Fork Basin, northwestern Wyoming. For more than a decade, this single specimen comprised the complete published record for T. brandti, until Smith et al. (2006) figured the P₄, M₁ and M₃ of the species from more complete specimens collected in the Clarks Fork and southern Bighorn basins. The upper dentition (P⁴-M³) was first described by Rose et al. (2011), along with the first dentary complete enough to characterize the lower dental formula (2.1.4.3), and the first postcranial elements attributed to T. brandti. In total, 31 dental specimens of T. brandti have been published to date. In contrast, the slightly younger North American species Bownomomys americanus (gen. nov., previously Teilhardina americana Bown, 1976, see Systematic paleontology) has been known for decades from numerous upper and lower dentitions that preserve the molars and posterior premolars through long time series documenting its hypothesized anagenetic evolution into B. crassidens (Bown and Rose, 1987). Similarly, T. belgica is represented by more than one hundred dental specimens that preserve many tooth positions (e.g., Gingerich, 1977). T. asiatica is represented by only four specimens (Ni et al., 2004, Rose et al., 2011)—though one of these includes a partial skull and associated lower dentaries. T. magnoliana is represented by 18 isolated cheek teeth (Beard, 2008), and ‘T. gingerichi’ is represented by only the holotype left dentary preserving P₃-M₁ (Rose et al., 2012).

Underestimating intraspecific phenotypic variation has the potential to distort and bias phylogenetic hypotheses and taxonomic diagnoses. This is frequently a challenge when evaluating fossil taxa known from very few specimens, but may play an even greater role when trying to distinguish closely-related species that have diverged recently—perhaps only within a few thousand years of the sampled populations. Improving the sample size and documenting morphological variation within species of Teilhardina is therefore important for distinguishing among members of this morphologically primitive, holarctically distributed genus and for establishing their relationships to the subsequent diversity of omomyiform taxa. Here we present a newly collected sample of 163 dental specimens of T. brandti from the southern Bighorn Basin that allows for description of previously unknown tooth positions and more thorough characterization of variation in phylogenetically and taxonomically relevant dental traits. These new morphological data are essential for testing the validity of the recently proposed species ‘T. gingerichi,’ also known from the PETM of the Bighorn Basin, Wyoming (Rose et al., 2012). The variation revealed by this sample also presents an opportunity to recontextualize T. brandti among other species of Teilhardina such as T. belgica and T. asiatica, and to rediagnose the anatomical features that differentiate these taxa. A cladistic analysis is performed to assess the effects these data have on proposed relationships between T. brandti and other omomyiforms, and their biogeographic implications for the dispersal of Teilhardina.