The Earliest Hominins: Sahelanthropus, Orrorin, and Ardipithecus

Until recently, the evolutionary events that surrounded the origin of the hominin lineage — which includes modern humans and our fossil relatives — were virtually unknown, and our phylogenetic relationship with living African apes was highly debated. Gorillas and chimpanzees were commonly regarded to be more closely related to each other due to their high degree of morphological and behavioral similarities, such as their shared mode of locomotion — knuckle-walking. But with the advent of molecular studies it has become clear that chimpanzees share a more recent common ancestor with humans, and are thus more closely related to us than they are to gorillas (e.g., Bailey 1993, Wildman et al. 2003). The similarities between the living African apes were thought to have been inherited from a common ancestor (=primitive features), implying that the earliest hominins and our last common ancestor shared with chimpanzees had features that were similar, morphologically and behaviorally, to the living African apes (Lovejoy 2009). With the discoveries of the earliest hominin species discussed below, it is now possible to critically examine these assumptions.

The chimpanzee-human divergence date has been estimated to be between 8 and 5 million years ago (MA) since the 1960s through immunologic and molecular techniques (e.g., Steiper & Young, 2006). Driven largely in part by these new genetic-based hypotheses, there have been intensive efforts by different teams over the last two decades to find and explore sediments that record this crucial time period for which we had virtually no fossil evidence. Their hard work and perseverance led to the discovery of several new genera and species of early hominins that are dated close to the estimated divergence dates for chimpanzees and humans. In 1994, Ardipithecus ramidus (ca. 4.4 Ma) was announced (White et al. 1994, 1995, WoldeGabriel et al. 1994) and soon after, even older hominins were discovered: Orrorin tugenensis (6.0-5.7 Ma, Pickford & Senut 2001, Senut et al. 2001, Sawada et al. 2002), Sahelanthropus tchadensis (7-6 Ma, Brunet et al. 2002, 2005, Vignaud et al. 2002), and Ardipithecus kadabba (5.8-5.2 Ma, Haile-Selassie 2001, WoldeGabriel et al. 2001).

These earliest hominins lack derived features found in later hominins, and their inclusion in the hominin lineage is largely based on a reduction in canine size, absence of the C/P3 honing complex, and the presence of morphological adaptations for habitual or obligate bipedality generally found in the postcranial skeleton, particularly in the pelvis and hindlimb. Bipedality is often considered to be the hallmark of hominins, and its presence in fossil species is often the key to their inclusion in the hominin clade. However, there is much contention and their status as hominins is contested by various researchers in the field (Wolpoff et al. 2002, 2006, Andrews & Harrison, 2005, Harrison, 2010, Wood & Harrison, 2011).

Sahelanthropus tchadensis from the site of Toros-Menalla, Chad (Figure 1), discovered by the Mission Paléoanthropologique Franco-Tchadienne (Brunet et al. 2002), may be the oldest hominin recovered thus far. However, the age of Sahelanthropus was first determined biochronologically (Vignaud et al. 2002). Unlike eastern African sites, Toros-Menalla lacks volcanic tuffs, which precludes the use of radiometric dating. Although an age range of 7–6 Ma is suggested for Sahelanthropus, cosmogenic nuclide dating using ¹⁰Be indicates that the sediments from which the fossils derive are at the older end of that range at 7.2–6.8 Ma (Lebatard et al. 2008).

Specimens of Sahelanthropus recovered include a complete, but crushed, cranium (TM 266-01-60-1), isolated mandibular fragments and dentition (Brunet et al. 2002, 2005). While the cranium possesses a host of primitive characters, it is argued to share derived features with later hominins that confirm its status as a member of the hominin clade. These include small canines worn at the tip, implicative of a reduced or absent C/P3 honing complex, and a short cranial base with a foramen magnum that is positioned anteriorly and orthogonal to the orbital plane, suggestive of an upright posture and habitual bipedality (Figure 2, Brunet et al. 2002, Guy et al. 2005, Zollikofer et al. 2005).

There is, however, much contention that Sahelanthropus is a hominin based on these traits. The determination of the small relative size of the canine is dependent on the supposition that the cranium is male as inferred from its large browridges. Various researchers have suggested that the link between browridge size and sex is tenuous — thus, a more parsimonious conclusion is that the cranium belongs to a female individual (Wolpoff et al. 2002, Andrews & Harrison 2005), specifically that of a female ape (and not a hominin at all) in whom canines are more likely to be worn at the tip (Wolpoff et al. 2002, 2006). They further argue that without diagnostic postcranial elements with clear adaptations for obligate bipedalism, such as the pelvis and femur, any inference made about the positional and locomotor behavior of Sahelanthropus is premature, because the taxonomic value and functional significance of shortened cranial base and foramen magnum position are unclear (Wolpoff et al. 2002, 2006, Andrews & Harrison 2005, Wood & Harrison 2011).

Specimens of Orrorin tugenensis were recovered from localities in the Lukeino Formation (6.0–5.7 Ma) of the Tugen Hills, Kenya (Figure 1) by the Kenya Paleontology Expedition. Among them were mandibular fragments, isolated teeth and a few postcranial elements (Senut et al. 2001). Generally speaking, the dentition is morphologically more similar to that of apes, particularly the canine, which is large, triangular and pointed at the apex (Senut et al. 2001). Because of this, arguments for the hominin status of Orrorin have coalesced around the likelihood of habitual bipedality in this species.

As one of the postcranial elements that shows diagnostic features of habitual bipedality, BAR 1002'00, a proximal femur, has been at the center of the debate. It shares several traits with later hominins that are interpreted as indicators of bipedality: elongated and antero-posteriorly compressed femoral neck, thicker cortex inferiorly than superiorly in the femoral neck, presence of an obturator externus groove, and well-developed gluteal tuberosity (Senut et al. 2001, Pickford et al. 2002, Galik et al. 2004). There has been particular emphasis on the asymmetric cortical bone distribution in the femoral neck, which is considered one of the key traits indicative of bipedality in Orrorin (Figure 3, Pickford et al. 2002, Galik et al. 2004). However, many researchers questioned the methodology employed and the quality of the CT scans used to demonstrate the inferiorly thicker femoral neck cortex of Orrorin (Ohman et al. 2005, White 2006), while others criticized the usefulness of the trait in general, regardless of its presence in Orrorin, as an indicator of habitual or obligate bipedality (Andrews & Harrison 2005). Senut et al. (2001) and Pickford et al. (2002) also claimed that the Orrorin femora were more human-like than Australopithecus species, and argued, based on this, that Orrorin led directly to Homo — thus, relegating Australopithecus to a side branch of the human evolutionary tree. However, recent morphometric study by Richmond & Jungers (2008) on BAR 1002'00 suggests that Orrorin is morphologically similar to Pliocene hominins, distinct from Homo and modern great apes, implying that it may have possessed a form of bipedality similar to that of Pliocene hominins.

Ardipithecus kaddaba is known from the late Miocene localities of the western margin of Middle Awash, Ethiopia (Figure 1). Specimens consist of mandibular fragments, isolated teeth and few postcranial elements recovered from the Asa Koma (5.54–5.77 Ma) and Kuseralee (ca. 5.2 Ma) Members of the Middle Awash study area (Haile-Selassie 2001, WoldeGabriel et al. 2001, Haile-Selassie et al. 2004, Renne et al. 2009). They were originally referred to as a subspecies of Ardipithecus ramidus (Haile-Selassie 2001), but were later elevated to a species based primarily on the primitive morphology of the C/P3 complex that implied the potential for some functional honing (Figure 4, Haile-Selassie et al. 2004, 2009). Among the postcranial elements recovered is a pedal proximal phalanx on which the proximal articular surface is dorsally canted (Figure 4). This feature is associated with toeing-off and is unique to bipeds (Latimer & Lovejoy 1990) — thus, linking Ar. kadabba to later hominins (Haile-Selassie 2001, Haile-Selassie et al. 2009). However, the fact that the pedal phalanx was recovered from the younger Kuseralee Member, while the dentition was from the older Asa Koma Member have prompted some doubts regarding its association to Ar. kadabba (Begun 2004).

Ardipithecus ramidus is suggested to be a descendent species of Ar. kadabba (Haile-Selassie et al. 2009). Specimens of Ar. ramidus have been recovered from the locality of Aramis (4.4 Ma) at Middle Awash, Ethiopia, by the Middle Awash project and from the locality of As Duma (4.51–4.32 Ma) at Gona, Ethiopia, by the Gona Palaeoanthropological Research Project (Figure 1, White et al. 1994, 2009a, WoldeGabriel et al. 1994, Semaw et al. 2005). Ardipithecus ramidus is unusual among the earliest hominin species in having a large sample size (110 specimens from Aramis) — the most complete and important of these is the recently described partial skeleton (ARA-VP-6/500), nicknamed "Ardi" (White et al. 2009).

The main traits that link Ar. ramidus to later hominins include: small, blunt canines, reduced canine size dimorphism, lack of a functional C/P3 honing complex, anteriorly positioned foramen magnum, and characters inferred to be indicative of bipedality, such as the presence of a greater sciatic notch, anterior inferior iliac spine, inferred lumbar lordosis, and dorsal canting of the pedal phalanx (Figure 5, White et al. 1994, 2009a, Semaw et al. 2005; Suwa et al. 2009a, 2009b, Lovejoy et al. 2009b, 2009c). However, "Ardi" also shows a number of primitive characters, particularly in the postcranium: fully opposable big toe, absence of longitudinal arch in the foot, relatively equal fore- to hind-limb lengths, and ape-like lower pelvis (Figure 5, White et al 2009a, Lovejoy et al. 2009a, 2009b, 2009c, 2009d). Thus, "Ardi" shows an unexpected mosaic of derived and primitive features that suggest it was a facultative biped, able to climb in the trees effectively by palmigrade quadrupedalism (Lovejoy et al. 2009d). This refutes the previous assumptions that the last common ancestor was chimpanzee-like, and instead suggests that living chimpanzees are highly specialized (Lovejoy et al. 2009d).

It has been noted, however, that Oreopithecus bambolii, a late Miocene ape from Italy, shows many of the characters that are used to denote the hominin status of Ar. ramidus. These include relatively small canines, reduction of C/P3 honing complex, anteriorly placed foramen magnum, and well-developed anterior inferior iliac spine. While other morphological features of Oreopithecus leave no doubt that it is an ape, the presence of the above characters suggests that interpretations of the earliest hominins may be confounded by homoplasies (Wood & Harrison 2011).

Ever since the discovery of Australopithecus africanus and the recovery of associated fauna indicative of open habitats, it has been posited that the origin of bipedality in our lineage had its roots in the expanding savannas of the Pliocene (Dart 1957). Subsequent discoveries of Australopithecus afarensis and associated fauna appeared to corroborate this hypothesis. However, the ecological contexts of the earliest hominins suggest that the link between the advent of bipedalism and the expansion of savanna and grassland habitats may be tenuous.

Current interpretations of the paleoenvironment suggest that Orrorin is associated with open woodland habitats with dense woodland or forest in the vicinity, possibly along lake margins (Pickford & Senut 2001). Sahelanthropus is likely found in a mosaic of environments, ranging from gallery forest at the edge of a lake, to savanna woodland, to open grassland, although there are indications that there was a dominance of shrub/bushland and grassy woodland habitats within the Chadian lake basin (Vignaud et al. 2002, Le Fur et al. 2009, Blondel et al. 2010). Ardipithecus kadabba is associated with riparian woodland and floodplain grassland along water margins (Su et al. 2009). Ardipithecus ramidus is found in closed woodland habitats with possible patches of forest at Aramis (White et al. 2009b, but see Cerling et al. 2010 for a different view) and associated with bushland and grassland habitats at Gona (Levin et al. 2008). Combined, these paleoenvironmental interpretations of the African latest Miocene and earliest Pliocene suggest that the beginnings of our lineage did not occur in open, semi-arid to arid habitat conditions, but rather in more closed and/or wet habitats. However, a definitive conclusion is difficult to draw at this time given the lack of detailed paleoecological reconstruction for Orrorin, the possibility that Sahelanthropus was found in more open habitats, the discordance in interpretation of the Aramis dataset, and the general paucity of late Miocene hominin-bearing sites in Africa.

While the hominin status of Sahelanthropus, Orrorin, and Ardipithecus species may be debated, their significance for understanding human origins and evolution is clear. They open a window into the very beginnings of our lineage, at the time when our ancestors diverged from those of our chimpanzee cousins. As more fossils of these species and their associated fauna are recovered and analyzed, we will learn more not only about their paleobiology, taxonomy, and phylogenetic relationships, but also about their environment, and how that may have influenced the evolution of our lineage.