Complete Fourth Metatarsal and Arches in the Foot of Australopithecus afarensis

See allHide authors and affiliations

Science  11 Feb 2011:
Vol. 331, Issue 6018, pp. 750-753
DOI: 10.1126/science.1201463

This article has a correction. Please see:


The transition to full-time terrestrial bipedality is a hallmark of human evolution. A key correlate of human bipedalism is the development of longitudinal and transverse arches of the foot that provide a rigid propulsive lever and critical shock absorption during striding bipedal gait. Evidence for arches in the earliest well-known Australopithecus species, A. afarensis, has long been debated. A complete fourth metatarsal of A. afarensis was recently discovered at Hadar, Ethiopia. It exhibits torsion of the head relative to the base, a direct correlate of a transverse arch in humans. The orientation of the proximal and distal ends of the bone reflects a longitudinal arch. Further, the deep, flat base and tarsal facets imply that its midfoot had no ape-like midtarsal break. These features show that the A. afarensis foot was functionally like that of modern humans and support the hypothesis that this species was a committed terrestrial biped.

Although Australopithecus afarensis was primarily a terrestrial biped, there continues to be debate over the nature of its bipedality and the extent to which its morphology represents a compromise between terrestrial bipedality and arboreal locomotion. One of the key adaptations to a human-like striding bipedal gait is the evolution of permanent transverse and longitudinal pedal arches (1, 2). The arches, supported by bone and soft tissue, provide an important mechanism for shock absorption during the stance phase of gait (3) and a rigid lever at heel-off, as well as permit flexibility during locomotion at different speeds and across irregular terrain (36). Muscles that in apes adduct the hallux (such as the m. adductor hallucis and m. fibularis longus), in humans primarily support the pedal arches (1, 5). Permanent plantar arches are a key component of human bipedal walking and running because they contribute to the rigidity of the foot and provide an enhanced mechanical advantage during the propulsive phase of gait (1, 79). Extant apes, in contrast, exhibit pronounced midtarsal dorsiflexion during heel-off as a result of a mobile midfoot, which permits flexibility for negotiating variably oriented arboreal substrates [(8, 9) and a recent review in (10)]. This break is greater in magnitude and is kinematically and anatomically distinct from the medial collapse seen in some humans (7, 11, 12). Therefore, determining the extent to which the foot of A. afarensis had permanent longitudinal and transverse pedal arches is key to deciphering the extent of its commitment to terrestrial bipedality.

Skeletal evidence for the presence of pedal arches in A. afarensis has been ambiguous, because key bones from the midfoot have been lacking. The talus (specimens AL 288-1 and AL 333-147) shows a distinct facet for plantar calcaneonavicular (13) and cubonavicular ligaments (9, 13, 14), which are indicative of a human-like medial longitudinal arch. However, unlike in humans, a groove is present for the m. fibularis longus tendon on the plantar surface of the ectocuneiform (AL 333-79) in A. afarensis, as seen in apes, perhaps related to the lack of a transverse arch. Dorsal inclination of the tarsal facets has been interpreted to suggest the lack of longitudinal arches in A. afarensis (11, 12), and the well-developed navicular tuberosity is argued to be a weight-bearing structure in these hominins (15, 16).

Here we describe AL 333-160, a complete, nearly perfectly preserved fourth metatarsal of A. afarensis from Hadar, Ethiopia (Fig. 1). This specimen was recovered from the Hadar locality AL 333 in 2000 during sieving of eroded Denen Dora 2 submember surface deposits of the Hadar Formation. Since 1975, these deposits at the 333 locality have yielded more than 250 hominin fossils that eroded from an in situ horizon dated to ~3.2 million years ago (17). We assign AL 333-160 to A. afarensis, the only hominin species in an assemblage of >370 hominin specimens so far recovered from the Hadar Formation (18). Other partial metatarsals attributed to A. afarensis are known from Hadar (19), but none is complete enough to address the question of pedal arches. The anatomy of a fifth metatarsal of A. africanus from Sterkfontein, South Africa (20, 21), is consistent with the presence of permanent pedal arches, but the fourth metatarsal is the key element along the lateral column of the foot that differs between apes and humans and is therefore the best test of the presence of permanent longitudinal and transverse arches in the foot.

Fig. 1

AL 333-160 left fourth metatarsal in dorsal, lateral, medial, plantar, and proximal views.

In AL 333-160, the metatarsal head is twisted laterally relative to the base, producing shaft torsion characteristic of modern humans (2) and later fossil hominins, including Homo habilis specimen OH 8 (22, 23) and the H. erectus foot bones from Dmanisi, Republic of Georgia (24). This torsion contrasts with the ape condition, in which the head and base exhibit minimal relative rotation (Fig. 2). Torsion allows the plantar surface of the metatarsal head to contact the ground in a foot with a strong skeletally supported transverse arch (2, 25, 26), an everted posture characteristic of a foot adapted for the modern human terminal-stance phase of gait, rather than the inverted foot postures of apes used in climbing. This degree of torsion of the AL 333-160 metatarsal demonstrates that a permanent bony transverse arch must have been present in the foot of A. afarensis.

Fig. 2

(A) Schematic representation of metatarsal proximal (rounded rectangles) and distal (ovals) ends seen in distal view in a human above and a chimpanzee below. In both species, the metatarsal heads are in contact with the substrate. Because of the arch, the proximal ends of the human metarsals are higher and situated in a transverse arch configuration (indicated by the dashed line). This results in axial torsion within the human fourth metatarsal of the head relative to the base, whereas this is not found in apes. [Modified from (23)] (B) Box plot of torsion values for chimpanzees, gorillas, humans (N = 10 individuals each), and AL 333-160, showing the distinct torsion in both hominins that is lacking in the apes. MT, metatarsal. Data are in table S1.

In AL 333-160, the diaphysis is angled plantarly, rather than dorsally, relative to the base, as in humans and H. habilis [OH 8; see (11)] and unlike in African apes (Fig. 3). This morphology further indicates a permanent longitudinally arched posture of the foot, because the fourth metatarsal makes an angle of about 8° to the ground in a normal human foot (5). The metatarsal head in AL 333-160 is flattened along the plantar portion of its articular surface, which faces distally relative to the diaphysis rather than being parallel to the diaphysis as in extant apes, forming a large plantar surface-diaphyseal angle (Fig. 3). This reflects the overall more extended posture of the metatarsophalangeal joints in the hominins (2, 5, 23).

Fig. 3

(A) Box plots of angular relations of the proximal and distal metatarsal ends to the diaphysis in chimpanzees, gorillas, humans, and AL 333-160. The proximal ends of hominin metatarsals are angled plantarly relative to the diaphysis, reflecting the average 8° of inclination of the metatarsal in normal arched posture, whereas that of the apes is oriented slightly dorsally. The flattened plantar portion of the hominin distal articular surface is inclined distally, also reflecting this posture and the habitual extension at this joint during bipedal locomotion, something also not seen in the apes who have this surface oriented directly plantarly. This distal plantar surface is also more distally oriented relative to the base in both hominins. In every case, AL 333-160 resembles humans only, strongly supporting the presence of arches in the A. afarensis foot. (B) Left fourth metatarsals of human, AL 333-160, chimpanzee, and gorilla in medial view, showing the orientation of bone ends and diaphysis. The blue arrows indicate the domed portion of the head. AL 333-160 resembles humans in having the doming along the dorsal articular margin, whereas the distal articular surface is domed more plantarly in the apes.

The AL 333-160 head exhibits another set of distinctive hominin apomorphies observed also in Ardipithecus ramidus (25), Australopithecus (14, 19, 27), and later hominins [reviews in (1, 22)]. It is domed dorsally in medial and lateral views (indicated by arrows in Fig. 3B), and there is a deep transverse gutter along the dorsal margin of the subchondral surface. In chimpanzees and gorillas, the domed portion of the head inclines plantarly, reflecting habitual loading in flexion. The hominin configuration seen in AL 333-160, and also in the AL 333-115 partial metatarsals (14, 19), would allow an increased range of dorsiflexion at the metatarsophalangeal joint as compared with apes, as well as habitual loading of the joint in extended postures that occur during the push-off and terminal phases of striding bipedal gait.

The lateral column of the human midfoot is relatively stiff, so that the mid- and hindfoot lift off the ground during gait simultaneously (1). In apes, however, dorsiflexion in the midfoot ensures that the heel leaves the substrate before the midfoot, a condition known as a “midtarsal break,” which can be up to 28° in magnitude (28). This dorsiflexion occurs primarily at the cuboid-metatarsal joints (10, 26, 29) and is distinct from the medial collapse in some human feet, which is far less pronounced and occurs to a variable degree at multiple joints (7, 11, 12). The transverse and longitudinal pedal arches and metatarsophalangeal dorsiflexion inferred from AL 333-160 signal an osteological pattern of midfoot stability and lateral foot rigidity unknown in the apes.

Dorsoplantar curvature of the lateral tarsometatarsal joint surfaces contributes to the distinctive midtarsal dorsiflexion in great apes (8, 10, 26, 29). These surfaces on the human proximal fourth and fifth metatarsals are flatter. The proximal articular surface of AL 333-160 is nearly flat (Fig. 4B), matching the mean of the modern human sample. Limited dorsiflexion at the lateral tarsometatarsal joints (10, 30), which would contribute to a relatively stiff lateral foot like that of modern humans, can be inferred for A. afarensis. AL 333-160 also has dorsoplantarly deep metatarsal bases, a condition also described for Ardipithecus ramidus (25) (Fig. 4B). This would limit dorsiflexion and plantarflexion at the lateral tarsometatarsal joints, additional evidence of a human-like relatively stiff lateral foot fundamentally different from that seen in apes.

Fig. 4

(A) Proximal ends of left fourth metatarsals in medial view, showing the dorsoplantar contour of the distal end. The box plot shows measured curvature, measured as maximum distance of the proximal joint surface from a line drawn between dorsal and plantar articular margins, expressed as a ratio to dorsoplantar length, following (11). Data are from (11). All hominins have relatively flat surfaces, rather than the convex profile of apes. (B) Proximal view of left fourth metatarsals, showing the dorsoplantarly expanded articular surface in hominins as compared with apes. The box plot of the ratio of dorsoplantar to mediolateral breadth shows the almost square proportion of apes, but the deep shape of the hominins. Data are in table S2. (C) Above, dorsal view of left fourth metatarsals, showing the articular facet for contact with the third metatarsal (vertical line) and the oblique articular facet for contact with the ectocuneiform in the hominins. Below, dorsal view of articulated cuboid, lateral, and medial cuneiforms and lateral metatarsals, showing the articular configuration of the lateral cuneiform with the third and fourth metatarsals. In apes, the cuneiform is directly medial to the cuboid and does not contact the fourth metatarsal. Both hominins have lateral cuneiform contact and an obliquely oriented facet on the fourth metatarsal for the cuneiform.

A rigid lateral foot in A. afarensis is further suggested by the orientation of the facet for the lateral cuneiform on the base of the AL 333-160 fourth metatarsal (Fig. 4C), mirroring the complementary facet seen on the lateral cuneiform (14, 19). In A. afarensis, as in modern humans, H. habilis [OH 8 (31)], and the Dmanisi H. erectus feet (24), the lateral cuneiform is elongated, extending distally past the cuboid, so that it articulates with the proximomedial corner of the fourth metatarsal at an obliquely oriented facet (14, 19). In apes, the lateral tarsometatarsal joints are aligned in the same coronal plane in such a way that the distal end of the lateral cuneiform is coplanar with the fourth tarsometatarsal joint, and the oblique facet on the fourth metatarsal for the lateral cuneiform is absent (Fig. 4C), a configuration that facilitates dorsiflexion at the tarsometatarsal joints (10). Thus, even if there was more calcaneocuboid mobility in A. afarensis than in modern humans (32, 33), this was evidently not the case for the lateral tarsometatarsal joints [see also (34)].

Most researchers conclude that the 3.6-million-year-old footprints in the Upper Laetolil Beds at Laetoli, Tanzania, evince a medial longitudinal arch [for example, (35, but see (36)]. Although A. afarensis is the only hominin species represented by fossil remains in these beds at Laetoli, one objection to this species having made the prints is the purported absence of the medial longitudinal arch in the Hadar foot (35, 37). The morphology of the AL 333-160 Hadar fourth metatarsal eliminates that objection.

The 4.4-million-year-old skeleton of Ardipithecus ramidus suggests that the transition to terrestrial bipedality occurred in the earliest hominins, while selection maintained adaptations in the foot for arboreal climbing and grasping (25). By at least 3.2 million years ago, the fundamental attributes of human pedal anatomy and function were in place. This includes the transformation of the first toe and associated musculature from a grasping structure to one designed for propulsion and shock absorption [review in (1)]. Evidence from the Hadar fourth metatarsal adds to this human-like portrait of permanent longitudinal and transverse bony arches in the sole of the foot. The evolutionary trajectory suggested by these fossil remains makes it unlikely that selection continued to favor substantial arboreal behaviors by the time of A. afarensis.

Supporting Online Material

References and Notes

  1. The authors thank the members of the Hadar Research Project (1990–2004) and the Hadar Paleoanthropology Field School (2007 and 2009) for their dedication and hard work, and the directors and staff of the National Museum of Ethiopia, Cleveland Museum of Natural History, and National Museums of Kenya for facilitating the analytical research reported here. Fieldwork permissions were kindly granted by the Authority for Research and Conservation of Cultural Heritage, Ethiopian Ministry of Culture and Tourism, and the Culture and Tourism Bureau of the Afar Regional State government. We thank J. DeSilva, J.M. Plavcan, and G. Schwartz for helpful comments. The analytical research was supported by NSF (grants NSF SBR-9601025 and NSF BCS-0333296), the University of Missouri Research Board and University of Missouri Research Council, and the Institute of Human Origins at Arizona State University.
View Abstract

Stay Connected to Science

Navigate This Article