Debates over the evolution of hominin bipedalism, a defining human characteristic, revolve around whether early bipeds walked more like humans, with energetically efficient extended hind limbs, or more like apes with flexed hind limbs. The 3.6 million year old hominin footprints at Laetoli, Tanzania represent the earliest direct evidence of hominin bipedalism. Determining the kinematics of Laetoli hominins will allow us to understand whether selection acted to decrease energy costs of bipedalism by 3.6 Ma.
Using an experimental design, we show that the Laetoli hominins walked with weight transfer most similar to the economical extended limb bipedalism of humans. Humans walked through a sand trackway using both extended limb bipedalism, and more flexed limb bipedalism. Footprint morphology from extended limb trials matches weight distribution patterns found in the Laetoli footprints.
These results provide us with the earliest direct evidence of kinematically human-like bipedalism currently known, and show that extended limb bipedalism evolved long before the appearance of the genus Homo. Since extended-limb bipedalism is more energetically economical than ape-like bipedalism, energy expenditure was likely an important selection pressure on hominin bipeds by 3.6 Ma.
Citation: Raichlen DA, Gordon AD, Harcourt-Smith WEH, Foster AD, Haas WR Jr (2010) Laetoli Footprints Preserve Earliest Direct Evidence of Human-Like Bipedal Biomechanics. PLoS ONE 5(3): e9769. https://doi.org/10.1371/journal.pone.0009769
Editor: Karen Rosenberg, University of Delaware, United States of America
Received: November 22, 2009; Accepted: February 28, 2010; Published: March 22, 2010
Copyright: © 2010 Raichlen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Funding was provided from the University of Arizona. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Ever since Darwin , bipedal walking has been considered the defining feature of the human lineage. However, how and why this unique form of locomotion evolved remains the subject of considerable debate. In particular, debates over the origins and evolution of bipedalism revolve around whether early bipeds walked with energetically economical human-like extended limb biomechanics, or with more costly ape-like bent-knee, bent-hip (BKBH) kinematics . If early hominins used a BKBH gait, then we must account for the persistence of an energetically costly form of bipedal walking until the evolution of the genus Homo. The Laetoli footprints may help resolve this debate, since they record the footsteps of at least two, and possibly three individuals who walked bipedally across wet ashfall approximately 3.6 million years ago , . These prints represent the earliest direct evidence of bipedalism in the fossil record, yet no study to date has demonstrated exactly how these hominins walked.
For decades, researchers have argued over whether the Laetoli hominins walked with a modern human-like extended limb gait , –, or a more ape-like form of bipedalism –. Most recently, the discovery of 1.5 Ma footprints from Kenya provides new evidence that the Laetoli prints were not completely modern in weight transfer and morphology . If true, then an energetically costly form of bipedalism evolved and persisted in early hominins until the evolution of the genus Homo . Alternatively, if the Laetoli prints were made by extended-limb bipeds, then, by 3.6 Ma, selection acted to reduce energy costs of locomotion in hominins.
Most previous studies of the Laetoli prints, however, were qualitative and did not test specific biomechanical hypotheses about early hominin gait. Kullmer et al.  performed a quantitative comparison of one Laetoli footprint (G1-36) with three footprints made by a single human subject walking with different postures. Their results suggest that Laetoli hominins did not walk like modern humans, however their sample size was very small (one human print for each posture studied) and there is not enough kinematic information to generalize their results to the Laetoli trackways. More recent quantitative work using geometric morphometrics techniques ,  may conflate differences in foot anatomy and shape with differences in locomotor biomechanics. Without using experimental analyses, it is difficult to determine which landmark differences between Laetoli and more recent fossil footprints are due simply to differences in foot morphology, and which are, in fact, due to differences in kinematics. Thus, a thorough experimental-based analysis of the Laetoli prints may resolve the debate over hominin biomechanics, and therefore help clarify the importance of selection for reduced energy costs of locomotion prior to the evolution of the genus Homo.
Here, we present the results of the first experimental analysis of footprints in a sample of humans walking with different gaits and compare our results to the Laetoli prints. Eight human subjects (mean body mass [SD] = 65.6 [6.1] kg) walked through a 5 m trackway filled with 15 cm of fine grained sand (0.075 mm–0.70 mm diameter). Although our subjects are likely heavier than the earlier hominins that generated the Laetoli prints, body mass does not appear to have an effect on footprint morphology . We compared footprints made by subjects walking with a normal, extended limb gait, and with a bent-knee, bent-hip (BKBH) ape-like gait at their preferred speeds with sand water content of 6–8% (see supplementary materials Text S1 for hind limb kinematic data from these experiments). These substrate conditions match those of Laetoli, which are described as similar to damp, fine to medium grained sand . We also examined the effects of increased speed and increased substrate moisture (10–12% water) on footprint morphology. We tested the hypothesis that a BKBH gait alters body weight transfer and produces significantly different footprint morphology than an extended limb gait. We predicted that Laetoli footprint morphology would match humans walking with one of these gaits. To test this prediction, we scanned each human print using a 3D laser scanner (Microsribe® MLX 6DOF digitizer with attached Microscan laser sensor system), and compared the maximum depths of the fore- (toe) and aft- (heel) sections of the prints. These values were calculated after leveling each print, since prints were made over a substrate that had a very small grade (mean grade = 1.49%±0.16%). We compared our experimental data with Laetoli footprint depths calculated from contour maps of the footprint trails  (Table S1). Additionally, we captured kinematic and kinetic data to determine how walking biomechanics influence footprint morphology (see supplementary materials Text S1).
Results and Discussion
There is a significant difference in the relative depth of the toes in BKBH compared to extended limb prints (Figs. 1 and 2; Table 1). BKBH footprints from humans walking at preferred speeds have toe depressions that are 76.69%±8.35% lower than the heel (calculated as toe depth as a percentage of heel depth; Table 1). When walking with an extended hind limb at preferred speeds, toe depressions are, on average, 22.36%±4.28% below the heel. Thus, BKBH gaits generate significantly greater toe relative to heel depths compared to extended limb gaits. Speed influences print morphology, with faster speeds leading to deeper toe depressions in both gaits, however between-gait differences are not significant (Table 1). Substrate moisture content also alters footprint morphology (Table 1). Wetter substrates lead to greater toe depths, regardless of gait, however, BKBH gaits have toe depths that are significantly greater than extended limb postures (Table 1).
Contours are 1 mm. A) Contour map of modern human footprint (Subject 6) walking with a normal, extended limb gait and side view of normal, extended limb footprint. B) Contour map of modern human footprint (Subject 6) walking with a BKBH gait and side view of BKBH print. C) Contour map of Laetoli footprint (G1-37) and side view of Laetoli footprint (G1-37). Note the difference in heel and toe depths between modern humans walking with extended and BKBH gaits. Laetoli has similar toe relative to heel depths as the modern human extended limb print.
Laetoli values were calculated using values from topographic maps. Note that the values for Laetoli fall within the range of human normal prints but outside of the range of human BKBH prints.
The difference in print morphology is related to a fundamental difference in body weight transfer between extended limb and BKBH gaits. The center of pressure (COP) is the point of ground reaction force application under the foot. As the COP travels from the heel to the toe during stance phase, its path, and the magnitude of ground force at any given moment, determines the amount of substrate displacement at a given position under the foot. During walking, as the COP shifts anterior to the metatarsal heads, the metatarso-phalangeal joint flexes, and because of the stiff longitudinal arch, the entire foot posterior to the metatarsal heads lifts off the ground . In a BKBH gait, the COP passes the metatarsal heads significantly earlier in the step (i.e., closer to heel-strike; Paired t-test p<0.001), and the ground reaction force impulse after the COP passes the metatarsal heads is significantly greater compared to extended limb gaits (Fig. 3; Paired t-test p<0.001). The larger impulse when the body is supported by the forefoot explains the increased toe-depth in BKBH footprints. Since most researchers agree that the Laetoli prints exhibit evidence of a longitudinal arch , , , , – (but see – and supplementary materials Text S1; Figs. S1, S2), an analysis of relative toe depths should provide unequivocal evidence of limb posture in these early hominins.
A) COP movement during bipedal walking. Fraction from touchdown (TD) is the mean time as a fraction of stance phase when the COP passes anterior to the metatarsal heads (error bars are SEMs). B) Late stance impulses during bipedal walking. Values are mean for the ground reaction force impulse after the COP passes anterior to the metatarsal heads (error bars are SEMs). Normal is extended limb walking, BKBH 1 is walking with a slightly flexed knee and hip, BKBH 2 is walking with a deeply flexed knee and hip.
The Laetoli prints have toe depths that are generally shallower than heel depths, however, the trackways were made on a very slight grade (3.12%±0.94%; see supplementary materials Text S1). After adjusting for this grade with the same procedures used for the experimental footprints (see supplementary materials Text S1 and above), mean toe depths for the G1 set of prints are generally equal to mean heel depths (0.11%±1.61% shallower than heels [0.00 mm±0.02 mm]), resembling weight transfer in a modern human-like extended limb gait more than a BKBH gait (Figs. 1 and 2; Supporting Table S1). In fact, while Laetoli proportional toe depths fall within the range of normal human patterns, Laetoli data fall outside of the range of human proportional toe depths made during BKBH trials (Fig. 2). We did not analyze the G2/3 set of footprints, since they were likely made by two individuals walking one behind the other, rendering their morphology less suitable for biomechanical determinations , , . There are two other caveats that must be considered in this analysis. First, bioturbation is evident in some of the Laetoli G1 prints , and could impact our results. However, much of the bioturbation occurred on the rims of the prints and does not greatly impact the internal morphology of most prints , . Second, the thickness of the substrate at Laetoli varies along the print trail , which may have an effect on footprint depths . However, the similarity of proportional depths across the G1 trail (Table S1) suggests that differences in substrate thickness did not impact our analysis. Finally, the Laetoli prints offer no evidence that these individuals were walking at fast speeds, since high speed walking also produces prints with much larger toe compared to heel depths for both gaits (see Table 1). Therefore, we conclude that the Laetoli hominins walked with an extended limb gait at speeds consistent with previous predictions (i.e., preferred or slow speeds) –, , –. The relative toe depths of the Laetoli prints show that, by 3.6 Ma, fully extended limb bipedal gait had evolved. Thus, our results provide the earliest unequivocal evidence of human-like bipedalism in the fossil record.
Hypotheses for the origins of bipedalism often focus on selection for energy economy in early hominins , . Energetic hypotheses are based on the reduced locomotor costs of humans compared to apes walking with BKBH gaits, and therefore, compared to ape-like pre-hominin ancestors . Human walking is inexpensive primarily because extended hind limb joints reduce external moments acting at the joints, and therefore, reduce the amount of muscle force required to support body weight , . In addition, extended-limb walking reduces joint reaction forces  and reduces total body heat loads compared to BKBH walking . By 3.6 Ma, hominins at Laetoli, Tanzania walked with modern human-like hind limb biomechanics, suggesting that selection for energetically economical bipedalism occurred prior to the evolution of the genus Homo. It is likely that reduced energy costs associated with extended limb bipedalism allowed early hominins to increase ranging distances during times of forest fragmentation  without enduring greatly increased energy costs.
While our results show that Laetoli hominins walked with human-like kinematics, we still cannot be sure of which hominin taxon made the footprints. Many researchers suggest that Australopithecus afarensis made the footprint trails , , , although this hypothesis is disputed by others based on differences between print morphology and fossilized foot remains , . If Au. afarensis did make the Laetoli footprints, then our results support the hypothesis that this species walked with relatively human-like hip and knee extension , , and that kinematically human-like bipedalism is compatible with adaptations for arboreality found throughout the australopith skeleton . Thus, settling the dispute over the taxonomic identification of the makers of the Laetoli prints will clarify debates surrounding fossil hominin post-cranial material and locomotor behavior , .
Finally, although our results clearly demonstrate that human-like bipedal kinematics had evolved by 3.6 Ma, the Laetoli prints cannot provide detailed information regarding the locomotor mechanics of earlier hominins. The recently published skeletal material attributed to Ardipithecus ramidus raises the hypothesis that, prior to 4.4 Ma, at least one lineage of hominins walked with kinematics that differed greatly from those of modern humans and other later hominins . However, until further functional morphological analyses are performed, it is difficult to assess the likely biomechanics of Ar. ramidus. Thus, based on the results of this study, kinematically human-like bipedalism clearly evolved within the first three to four million years of hominin evolution. However, we are left with two possible scenarios for the origins and evolution of bipedalism. First, if a detailed functional analysis supports the hypothesis that Ar. ramidus was a habitual bipedal hominin that walked with flexed lower limbs, then energetic selection pressures likely became strongest after the origins of bipedalism, but prior to 3.6 Ma. This finding would not necessarily rule out the energy hypothesis for the origins of bipedalism , but would suggest that early bipeds were less energetically economical than modern humans. Second, if Ar. ramidus was not a bipedal hominin, then the earliest bipeds may have walked with extended limb joints, suggesting that early bipedalism was as energetically economical as that of later hominins, including those that made the Laetoli footprints. Testing these hypotheses requires more detailed fossil evidence from the earliest hominins. However, our analysis of the Laetoli prints refines the timing of the evolution of human-like bipedal mechanics in the fossil record. Future analyses of fossil remains, and future fossil discoveries, will no doubt improve our understanding of just how early human-like bipedal mechanics evolved, and therefore, help us determine the importance of selection for energy economy during the early evolution of bipedal walking.
Eight subjects participated in this study (see Table S2). All methods were approved by the University of Arizona Human Subjects Committee and subjects gave their written informed consent prior to participation.
All subjects were asked to walk at their preferred walking speed through a sand trackway (see Table S3). After making two passes down the walkway at their preferred speed, subjects were asked to walk with a bent-knee, bent-hip (BKBH) gait at their preferred speed. Subjects also repeated the BKBH trial twice. The trackway was 15 cm deep and 5 m long. It was filled with mostly fine sand, with coarser grained sand confined to the beginning and end of the trackway. Grain sizes ranged from <0.074 mm to 0.7 mm in diameter. After walking at their preferred speeds, a subset of the sample was asked to walk at a fast speed using both gaits. A different subset of the sample was asked to walk in both gaits at their preferred speeds after the moisture content of the sand was increased by ∼4%. Moisture content of sand was measured using a HydroSense (Campbell Scientific CD 620) soil moisture system. For low moisture trials, moisture content was 6–8%. For high moisture trials, moisture content was 10–12%. To determine moisture content, the moisture probe was inserted into the footprint after each scan was taken.
Prior to trackway trials, reflective markers were affixed to the hip, knee, ankle, heel, hallux, 1st metatarsal head, and shoulder. Subjects were filmed walking through the trackway using a six-camera Vicon motion analysis system (200 Hz). Maximum, minimum, and average joint angles were calculated for the hip and knee during all trials (Table S4).
After each pass through the trackway, footprints were digitized using a 3-dimensional laser scanner (Microscribe MLX 6DOF digitizer with attached Micorscan laser sensor system). At least two prints (left and right) for each trial were scanned. For each scanned print, four points were selected from the sand around the print: two points medial and lateral to the toes, and two points medial and lateral to the heel. A plane representing the sand substrate was generated based on those four points using a least-squares fit in Microscan Tools. The 3D coordinates of the print and plane were then imported into R where the plane and print were reoriented such that the plane was level. Finally, the depths of the deepest point in the heel and in the toe were extracted from the leveled print, and proportional toe depth was calculated as a fraction of the depth of the heel.
A second sample of subjects (n = 6) were recruited to determine the kinetics of BKBH walking. Subjects walked over a forceplate (AMTI; 4000 Hz) and were simultaneously filmed with a six-camera Vicon motion analysis system (200 Hz). Prior to walking trials, reflective markers were affixed to the hip, knee, ankle, heel, hallux, 1st metatarsal head and shoulder. Subjects walked at preferred speeds using normal extended limb bipedalism as well as two BKBH trials (one with light and one with deep knee and hip flexion; Table S5).
Center of pressure data were collected from force plate recordings for each trial. Combing kinematic and kinetic data, we calculated the time of heel strike (touchdown) and the time (relative to heel strike) that the COP passed anterior to the metatarsal heads. We also calculated the total impulse of the ground reaction force from the time the COP passed the metatarsal heads to toe-off. These values provide a biomechanical measure of how toe depths are generated. For more information on methods see supplementary materials Text S1.
Sample size information for footprint analyses
(0.03 MB DOC)
Speed and joint angles for trackway trials. Values are means (SEM) for all subjects.
(0.03 MB DOC)
Hip and knee angles for force plate trials.
(0.03 MB DOC)
Proximal pressure ridge in a normal human footprint.
(3.70 MB TIF)
We thank William Jungers, Daniel Lieberman, and Herman Pontzer for helpful discussions of this project. Vija Garcia, Justin Higgins, Alicia Waltz, Stephanie Reyes, and David Zahn provided assistance during data collection. We thank Mitchell Pavao-Zuckerman for generously providing us with equipment to measure soil moisture content. We are grateful to Lars Fogelin for assistance in trackway construction. We also thank all of our subjects for their time and effort.
Conceived and designed the experiments: DR ADG WEHHS. Performed the experiments: DR ADG. Analyzed the data: DR ADG. Contributed reagents/materials/analysis tools: ADG WEHHS AF WRHJ. Wrote the paper: DR ADG WEHHS AF.
- 1. Darwin C (1871) The decent of man, and selection in relation to sex. London: John Murray (Reprinted in 1981 by Princeton University).
- 2. Stern J (2000) Climbing to the top: a personal memoir of Australopithecus afarensis. Evolutionary Anthropology 9: 113–133.
- 3. Leakey MD, Hay RL (1979) Pliocene footprints in the Laetoli beds at Laetoli, northern Tanzania. Nature 278: 317–323.
- 4. Leakey MD (1981) Tracks and tools. Philisophical Transactions of the Royal Society of London 292: 95–102.
- 5. Day MH, Wickens EH (1980) Laetoli pliocene hominid footprints and bipedalism. Nature 286: 385–387.
- 6. White TD (1980) Evolutionary implications of pliocene hominid footprints. Science 208: 175–176.
- 7. White TD, Suwa G (1987) Hominid footprints at Laetoli: Facts and Interpretations. Am J Phys Anthropol 72: 485–514.
- 8. Tuttle R (1985) Ape footrpints and Laetoli impressions: A response to the SUNY claims. In: Tobias P, editor. Hominid evolution: Past, present, and future. New York: Alan R. Liss. pp. 129–133.
- 9. Tuttle R (1987) Kinesiological inferences and evolutionary implications from Laetoli bipedal trails G-1, G-2/3, and A. In: Leakey MD, Harris J, editors. Laetoli A Pliocene site in northern Tanzania. Oxford UK: Clarendon Press. pp. 503–523.
- 10. Tuttle R, Webb D, Tuttle N (1991) Laetoli footprint trails and the evolution of hominid bipedalism. In: Coppens Y, Senut B, editors. Origine(s) de la bipedie chez les hominides. Paris: Cahiers de Paleioanthropologie, Editions du CNRS. pp. 203–218.
- 11. Stern JT, Susman RL (1983) The locomotor anatomy of Australopithecus afarensis. Am J Phys Anthropol 60: 279–317.
- 12. Deloison Y (1992) Emprientes de pas a Laetoli (Tanzanie). CR Acad Sci Paris 315: 103–109.
- 13. Meldrum D (2004) Midfoot flexibility, fossil footprints, and sasquatch steps: New perspectives on the evolution of bipedalism. Journal of Scientific Exploration 18: 65–79.
- 14. Meldrum D (2007) Renewed perspective on the Laetoli trackways: The earliest hominid footprints. In: Lucas , Spielman , Lockley , editors. Cenozoic vertebrate tracks and traces: New Mexico Museum of Natural History and Science Bulletin.
- 15. Bennett M, Harris J, Richmond B, Brain D, Mbua E, et al. (2009) Early hominin foot morphology based on 1.5-million-year-old footprints from Ileret, Kenya. Science 323: 1197–1201.
- 16. Sockol MD, Raichlen DA, Pontzer H (2007) Chimpanzee locomotor energetics and the origin of human bipedalism. Proceeding of the National Academy of Science 134: 12265–12269.
- 17. Kullmer O, Schrenk F, Dorrhoffer B (2003) High resolution 3D-image analysis of ape, hominid and human footprints. Courier Forsch Senkenberg 243: 85–91.
- 18. Berge C, Penin X, Pelle E (2006) New interpretation of Laetoli footprints using an experimental apprach and procrustes analysis: Preliminary results. CR Palevol 5: 561–569.
- 19. D'Aout K, Meert L, Van Gheluwe B, De Clercq D, Aerts P (in press) Experimentally generated footprints in sand: Analysis and consequences for the interpretation of fossil and forensic footprints. Am J Phys Anthropol. DOI:https://doi.org/10.1002/ajpa.21169.
- 20. Hay RL, Leakey MD (1982) The fossil footprints of Laetoli. Sci Am 246: 50–57.
- 21. Leakey MD, Harris JM (1987) Laetoli, A Pliocene Site in Northern Tanzania. Oxford: Oxford University Press.
- 22. Imhauser C, Siegler S, Abidi N, Frankel D (2004) The effect of posterior tibialis tendon dysfunction on the plantar pressure characteristics and the kinematics of the arch and hindfoot. Clinical Biomechanics 19: 161–169.
- 23. Feibel CS, Agnew N, Latimer B, Demas M, Marshall F, et al. (2005) The Laetoli hominid footprints - A preliminary report on the conservation and scientific study. Evolutionary Anthropology 4: 149–154.
- 24. Schmid P (2004) Functional interpretation of the Laetoli footprints. In: Meldrum DJ, Hilton C, editors. From biped to strider: The emergence of modern human walking, running, and resource transport. New York: Kluwer Academic/Plenum Publishing. pp. 49–61.
- 25. Tuttle RH (1996) The Laetoli hominid G footprints. Where do they stand today? Kaupia 6: 97–102.
- 26. Tuttle RH, Webb D, Weidl E, Baksh M (1990) Further progress on the Laetoli trails. Journal of Archaeological Science 17: 347–362.
- 27. Robbins LM (1987) Hominid footprints from site G. In: Leakey MD, Harris JM, editors. Laetoli: A pliocene site in northern Tanzania. Oxford: Clarendon Press. pp. 497–502.
- 28. Hay RL (1987) Geology of the Laetoli area. In: Leakey MD, Harris JM, editors. Laetoli: A pliocene site in northern Tanzania. Oxford: Oxford University Press. pp. 23–47.
- 29. Charteris J, Wall JC, Nottrodt JW (1982) Pliocene homind gait: New interpretations based on available footprint data from Laetoli. Am J Phys Anthropol 58: 133–144.
- 30. Alexander RM (1984) Stride lengths and speed for adults, children, and fossil hominids. Am J Phys Anthropol 63: 23–27.
- 31. Sellers W, Cain G, Wang W, Crompton RH (2005) Stride length, speed and energy costs in walking of Australopithecus afarensis: Using evolutionary robotics to predict locomotion of early human ancestors. Journal of the Royal Society Interface 2: 431–441.
- 32. Raichlen D, Pontzer H, Sockol MD (2008) The Laetoli footprints and early hominin locomotor kinematics. J Hum Evol 54: 112–117.
- 33. Rodman PS, McHenry HM (1980) Bioenergetics and the origin of hominid bipedalism. Am J Phys Anthropol 52: 103–106.
- 34. Pontzer H, Raichlen D, Sockol MD (2009) The metabolic cost of walking in humans, chimpanzees and early hominins. J Hum Evol 56: 43–54.
- 35. Carey TS, Crompton RH (2005) The metabolic costs of ‘bent-hip, bent-knee’ walking in humans. J Hum Evol 48: 25–44.
- 36. Crompton RH, Li Y, Wang W, Gunther M, Savage R (1998) The mechanical effectiveness of erect and ‘bent-knee, bent-hip’ bipedal walking in Australopithecus afarensis. J Hum Evol 35: 55–74.
- 37. Cerling T, Harris JM, MacFadden B (1997) Global vegetation change through the Miocene/Pliocene boundart. Nature 389: 153–158.
- 38. Harcourt-Smith WEH (2005) Did Australopithecus afarensis make the Laetoli footprint trail? New insights into an old problem. American Journal of Physical Anthropology Supplement S40: 116.
- 39. Latimer B, Lovejoy CO (1989) The calcaneus of Australopithecus afarensis and its implications for the evolution of bipedality. Am J Phys Anthropol 78: 369–386.
- 40. Ward CV (2002) Interpreting the posture and locomotion of Australopithecus afarensis: Where do we stand? Yearbook of Physical Anthropology 45: 185–215.
- 41. Lovejoy CO, Suwa G, Spurlock L, Asfaw B, White TD (2009) The pelvis and femur of Ardipithecus ramidus: The emergence of upright walking. Science 326: 71e71–71e76.