Comparative description and taxonomic affinity of 3.7-million-year-old hominin mandibles from Woranso-Mille (Ethiopia)

Hominin fossil discoveries during the last three decades have redefined our understanding of the tempo and mode of human evolution during the Plio-Pleistocene (Brunet et al., 1996, 2002; Leakey et al., 2001; Alemseged et al., 2006, 2020; White et al., 2006; Berger et al., 2010, 2015; Haile-Selassie et al., 2010a, 2010b, 2012, 2015, 2019; Villmoare et al., 2015; Herries et al., 2020). The discovery of the earliest known fossil of the Homo lineage at 2.8 Ma, in the Afar region of Ethiopia (Villmoare et al., 2015; DiMaggio et al., 2015), the recognition of different hominin species contemporaneous with Australopithecus afarensis from the mid-Pliocene period in eastern and central Africa (Australopithecus bahrelghazali, Kenyanthropus platyops, Australopithecus deyiremeda; Brunet et al., 1996; Leakey et al., 2001; Spoor et al., 2010, 2016; Haile-Selassie et al., 2012, 2015), and recent fossil discoveries of previously known species such as Australopithecus anamensis (Haile-Selassie et al., 2019), have challenged several long-held ideas and hypotheses about early hominin taxonomic diversity and phylogenetic relationships. These advances underscore the value of each newly discovered fossil for increasing our understanding of human evolutionary history. Although new areas in the Afar region, the Turkana Basin (Kenya), and other parts of Africa have been yielding important remains of early hominins (see Brown et al., 2013), Woranso-Mille, a site located in the north-central part of the Afar region of Ethiopia, has become one of the most prolific mid-Pliocene sources: in the last 15 years, more than 200 hominin specimens of Pliocene age have been recovered from the site.

Woranso-Mille (Fig. 1) presents the strongest evidence for the contemporaneous presence of at least one non-Au. afarensis hominin species (e.g., Au. deyiremeda or a species represented by the Burtele foot) between 3.5 and 3.3 Ma (Haile-Selassie et al., 2012; Wood and Boyle, 2016), whereas the biological validity of other species named from the same period (Au. deyiremeda, Au. bahrelghazali, and K. platyops; Brunet et al., 1996; Leakey et al., 2001; Haile-Selassie et al., 2015) has been contested on various grounds (White, 2003; Kimbel et al., 2004; Wood and Boyle, 2016; but see Guy et al., 2008; Spoor et al., 2010; Haile-Selassie et al., 2016a). The Woranso-Mille fossils show that Au. afarensis and Au. deyiremeda coexisted in close geographic and temporal proximity at 3.5–3.3 Ma (Haile-Selassie et al., 2016a; Melillo et al., 2021). The 3.4 Ma Burtele partial foot from Woranso-Mille also indicated the presence of a bipedal locomotor adaptation different from that of the contemporaneous Au. afarensis (Haile-Selassie et al., 2012). Despite the close temporal and spatial provenience of the foot to specimens assigned to Au. deyiremeda (Haile-Selassie et al., 2015), it has not been assigned to any species yet, pending recovery of more complete craniodental specimens that are taxonomically informative and whose provenience could be clearly associated with the foot. Woranso-Mille fossiliferous sediments older than 3.5 Ma have yielded a 3.6-million-year-old partial skeleton and juvenile mandibles of Au. afarensis (Haile-Selassie et al., 2010a; Haile-Selassie and Ryan, 2019) and more recently a 3.8-million-year-old cranium of Au. anamensis (Haile-Selassie et al., 2019; Saylor et al., 2019). The Au. anamensis cranium (MRD-VP-1/1) not only expanded the previously known temporal range of the species (4.2–3.9 Ma; Leakey et al., 1995; Ward et al., 2001) but also provided a more complete understanding of the craniofacial anatomy of the species. Furthermore, it challenged the Au. anamensis–Au. afarensis anagenesis hypothesis by showing that the two species may have overlapped in time (Haile-Selassie et al., 2019).

The possible presence of Au. anamensis at Woranso-Mille was first hinted at by the two ca. 3.76 Ma partial mandibles (Haile-Selassie, 2010) that are described in detail here. Preliminary observations indicated that they retained a mosaic of dental and mandibular features intermediate between those of Au. anamensis and Au. afarensis. Given the presence of some features that are seen only in Au. anamensis, however, Haile-Selassie (2010) argued that these mandibles might be best assigned to Au. anamensis. Here, we provide a detailed comparative description of these specimens, particularly because more parts of one of the mandibles were recovered as a result of continued fieldwork, and newly assess their taxonomic affinity. This investigation will have significant implications for our understanding of the temporal and spatial distribution and dentognathic morphological variation in the earliest Australopithecus species in eastern Africa.

The left half of MSD-VP-5/16 was found in 2006 by Alemayehu Asfaw and the right half was found by a local Afar worker (Ali Kadir) in 2018, ca. 80 m south of the initial discovery (Figs. 2 and 3). The right half of the mandible joined the left half along the anterior corpus and symphyseal region. The corpus is well preserved with minimal to no distortion on its preserved parts on both sides. It does not have any sign of weathering or carnivore damage. Anteriorly, the adjoining break runs obliquely from the distal wall of the left I2 at the alveolar margin to the base of the anterior corpus at the midline. The corpus on the left side is almost intact, extending posteriorly past the origin of the ramus, which is broken obliquely toward the gonial angle. Most of the ramus is missing on both sides except for the anteroinferior portion of the masseteric fossa at the M3 level on the left side. The corpus base is partially preserved on both sides and in good condition, except for hairline cracks running along the length of the preserved corpus bases. Also on the lateral face of the left corpus base, there is another hairline crack that runs from the level of the C1 posteriorly to the level of the mesial M2. A small area of cortical bone (12.3 mm anteroposteriorly and 3.1 mm mediolaterally) is also flaked from the preserved distal terminus of the left corpus base at the M2 level. The symphyseal region also shows some cracks that do not impact this region's morphology. A vertical crack runs down across the entire length of the median ridge formed by the incisal juga. Another crack, where the two halves joined, traverses obliquely from the base of the left I2 to the midline at the base. On the right side, the corpus is broken vertically, but irregularly, posterior to the M1. The preserved right lateral corpus shows a small triangular area of laterally deflected cortical bone close to the base at the P4–M1 level. There is also a faint crack running from the alveolar margin of the P4 all the way to the base, traversing right through the mental foramen and extending to the corpus base. This crack continues to the medial side, running anterosuperiorly, and ends at the genioglossal fossa. The corpus base, as on the left side, has a hairline crack that traverses the entire length of the corpus base but causes no distortion. Also on the right side, the digastric fossa is depressed because of cortical bone loss at its center. The P4 on the right side, the M1s on both sides, and the left M2 are preserved intact. A small piece of cortical bone is flaked from the distolingual corner of the left M2 partially exposing its distal root. A small area of enamel is chipped from the apex of the left M1 protoconid and a hairline crack runs across the occlusal surface of the hypoconid. The right M1 protoconid was broken from the crown and its mesial half was recovered from a sieving operation. This molar also has a hairline crack that runs centrally from the central fovea to the distal marginal ridge. All of the anterior tooth crowns are missing, most of them broken at the level of the alveolar margin. The left M3 crypt is partially preserved although its medial wall is broken. Its overall size and C1 root length indicate that it probably belonged to a female individual while the unerupted M3 with partially formed roots allows for an estimation of its developmental status (see section 3.1 for further details).

MSD-VP-5/50 is a left mandible preserving the P3 to the M3 found in 2009 by Woganu Amerga. It was found broken into numerous small pieces most of which joined (Figs. 4 and 5). The anterior corpus is partially preserved below the I2 alveolus and at its base, the bone extends close to the midline. The anterior break is not fresh and the edges are rounded. Most of the inferior half of the ramus is preserved with some areas missing in the middle. The posteroinferior corner of the ramus above the gonial angle is also missing. The preserved corpus does not show substantial biotic or abiotic alterations and visible cracks are mostly from joined surfaces. The lateral and lingual alveolar margins are damaged to variable degrees. On the lateral side, the alveolar margin is unevenly chipped from C1 to M1 while it is preserved intact at both the distal M2 and distal M3 regions. On the lingual side, however, the margin is well preserved from P4 to mesial M2. Based on its overall corpus size, C1 root robusticity, and length (judged from its preserved socket), it probably belonged to a male, while the amount of wear on its molars indicates that it was an older individual (Fig. 5).

Woranso-Mille (Fig. 1) is located in the central Afar region of Ethiopia north of Hadar where Au. afarensis is best known from 3.40 to 2.95 Ma (Kimbel et al., 2004; Campisano and Feibel, 2007, 2008; Kimbel and Delezene, 2009). Geological and paleontological work at Woranso-Mille largely concentrated on the ca. 180-m-thick fossiliferous and sedimentary strata exposed in the modern Mille River catchment. These strata are radiometrically dated to 3.8–3.2 Ma, with the faunal assemblages divided into five age groups based on dated marker tuffs and paleomagnetic boundaries (3.8–3.66, 3.66–3.57, 3.57–3.47, 3.47–3.33, and 3.33–3.2 Ma). Mesgid Dora locality 5 (MSD-VP-5), one of nine localities in the MSD collection area, is among the older localities in the 3.8–3.66 Ma time interval (Fig. 1). The general stratigraphy of locality MSD-VP-5 has been described previously, including 40Ar/39Ar dating and tephrochemical correlation of volcanic tuffs and mapping of volcanic tuffs and channel sandstone bodies (Deino et al., 2010; Saylor et al., 2016). The fossil specimens described here were collected ∼250 m apart, from the top surface of a felsic lapillistone tuff that is part of the Mille Tuff sequence (Fig. 1). In most locations the tuffs in the Mille sequence are airfall deposits (Saylor et al., 2016, 2019), but some have been fluvially reworked into ∼5 m of interbedded and cross-bedded tuffaceous sandstone, siltstone and discontinuous tuff that lie between the fossil collection horizon and an overlying bimodal tuff composed of basaltic and felsic pumice (Fig. 1). Ages for the Mille Tuff sequence (3.76 ± 0.02 Ma) and the Araskimiro Tuff (3.77 ± 0.04 Ma), a felsic pumice tuff above the bimodal tuff, are statistically indistinguishable (Deino et al., 2010; Saylor et al., 2016).

The sequence of rhyolitic and basaltic lapillistone tuffs and tuffaceous sedimentary rocks that contains the fossil collection horizon is overlain by coarse channelized sandstone followed by the Waki Tuff (3.664 ± 0.016 Ma), an extrabasinal vitric ash that has chronological and geochemical characteristics compatible with the Lomogol Tuff in the Omo-Turkana Basin (Saylor et al., 2016). The pumice tuffs, sandstone, and the Waki Tuff have been traced from the MSD-VP-5/50 discovery location to within 50 m of the MSD-VP-5/16 discovery (Saylor et al., 2016), though in the near proximity to the site itself these beds are missing because of subsequent incision by an intraformational fluvial channel (Fig. 6). Conglomeratic channel sandstones contain diverse basaltic and rhyolitic clasts up to 5 cm across and form northwest-southeast to northeast-southwest oriented curvilinear bodies that are 20–30 m across and on the order of several hundred meters in length. These channel sandstones were deposited in a broad paleo-valley that is ∼600 m across and oriented northwest-southeast. The Mesgid Dora Tuff, which lies stratigraphically above the Waki Tuff, drapes more than 15 m of relief along this paleo-valley and varies from airfall deposits that are as much as ∼10 m above the Waki Tuff to fluvially reworked channel deposits directly overlying the Mille Tuff Sequence (Saylor et al., 2016). Beds of limestone and platy mudstones formed in shallow water bodies within the paleo-valley. Lack of well-developed paleosols is consistent with a quickly aggrading depositional system and short-lived land surfaces. The maximum age of MSD-VP-5/16 and MSD-VP-5/50 is established by the Mille Tuff Sequence (3.76 ± 0.02 Ma), and the minimum age by the Waki Tuff (3.664 ± 0.016 Ma), though the specimens are more likely closer to the maximum age because they were recovered from immediately above the Mille Tuff Sequence (Fig. 1).

The faunal assemblage from MSD-VP-5 is similar to that from contemporaneous and slightly younger Woranso-Mille localities at Aralee Issie (Haile-Selassie et al., 2010b). Previous multiproxy paleoecological studies of the faunal assemblage at MSD (3.76–3.57 Ma) indicated heterogeneous habitats with dense vegetation along a paleo-river grading into more open habitats away from the riparian zone (Curran and Haile-Selassie, 2016). Su and Haile-Selassie (2022) further showed that the nonhominin faunal community structure at MSDC (a fossil collection area that includes MSD-VP-5 and two other nearby localities dated to between 3.76 and 3.66 Ma) indicates heterogeneous habitats while it appears that woodland habitats were predominant.

留言 (0)

沒有登入
gif