lable at ScienceDirect

Journal of Human Evolution 59 (2010) 608e619

Contents lists avai

Journal of Human Evolution

journal homepage: www.elsevier .com/locate/ jhevol

Comparative in vivo forefoot kinematics of Homo sapiens and Pan paniscus

Nicole L. Griffin a,*, Kristiaan D’Août b,c, Brian Richmond d,e, Adam Gordon f, Peter Aerts b,g

aDepartment of Evolutionary Anthropology, Duke University, P.O. Box 90383 Science Drive Durham, NC, 27708-0383, USAbDepartment of Biology, University of Antwerp, BelgiumcCentre for Research and Conservation, Royal Zoological Society of Antwerp, BelgiumdCenter for the Advanced Study of Hominid Paleobiology, Department of Anthropology, The George Washington University, USAeHuman Origins Program, National Museum of Natural History, Smithsonian Institution, USAfDepartment of Anthropology, University at Albany e SUNY, USAgDepartment of Movement and Sports Sciences, University of Ghent, Belgium

a r t i c l e i n f o

Article history:Received 10 September 2009Accepted 10 July 2010

Keywords:Great apeHalluxHumanJoint excursionMetatarsophalangeal jointTerrestrial locomotion

* Corresponding author.E-mail address: nicole.griffin@duke.edu (N.L. Griffi

0047-2484/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.jhevol.2010.07.017

a b s t r a c t

The human metatarsophalangeal joints play a key role in weight transmission and propulsion duringbipedal gait, but at present, the identification of when a habitual, human-like metatarsi-fulcrimatingmechanism first appeared in the fossil record is debated. Part of this debate can be attributed to theabsence of certain detailed quantitative data distinguishing human and great ape forefoot form andfunction.

The aim of this study is to quantitatively test previous observations that human metatarsophalangealjoints exhibit greater amounts of dorsal excursion (i.e., dorsiflexion) than those of Pan at the terminalstance phase of terrestrial locomotion. Video recordings were made in order to measure sagittalexcursions of the medial metatarsophalangeal joints in habitually shod/unshod adult humans and adultbonobos (Pan paniscus). Results indicate that the human first and second metatarsophalangeal jointsusually dorsiflex more than those of bonobos. When timing of maximum excursion of the first meta-tarsophalangeal joint is coupled with existing plantar pressure data, the unique role of the humanforefoot as a key site of leverage and weight transmission is highlighted. These results supporthypotheses that significant joint functional differences between great apes and humans during gaitunderlie taxonomic distinctions in trabecular bone architecture of the forefoot.

� 2010 Elsevier Ltd. All rights reserved.

Introduction

Modern human bipedal gait is unique in many ways, includinga long stride with an extended hip that contributes to walkingefficiency (Alexander, 2004; Sockol et al., 2007; Crompton et al.,2008). Part of this unique gait complex involves a forefoot formand function that differs from the closest living relatives of modernhumans, the great apes (Elftman and Manter, 1935; Morton, 1964;Bojsen-Møller, 1979; Wunderlich, 1999; Vereecke et al., 2003).Comparisons between great ape and modern human hard and softtissue characteristics of the foot clearly highlight the specializationof the modern human metatarsophalangeal joints for habitualbipedality.

Bipedal adaptations that distinguish the modern human fore-foot from that of the great ape forefoot include an adducted 1stmetatarsal (MT) that is substantially larger than the lateral MTs,

n).

All rights reserved.

distal articular regions of the MTs that are oriented in a way toallow greater metatarsophalangeal joint excursion in the dorso-plantar plane, and relatively short pedal phalanges (Morton, 1964;Susman, 1983; Rolian et al., 2009). In addition, the modern humanfoot is characterized by a thick plantar aponeurosis, a ligamentousband that originates from the calcaneus and through a complexnetwork, attaches to each proximal phalanx (PP) (Hicks, 1954;Bojsen-Møller and Lamoreux, 1979; Erdemir et al., 2004). Theplantar aponeurosis maintains the longitudinal arch of the footduring gait (Hicks, 1954; Caravaggi et al., 2009), transmits forcesbetween the hindfoot and the forefoot during the stance phase(Erdemir et al., 2004), cushions against the high ground reactionforces that occur at the ball of the foot during the latter part of thestance phase (Bojsen-Møller and Lamoreux, 1979), and acts as anenergy saving mechanism (Ker et al., 1987). Many researchers havenoted that great apes lack a well-developed plantar aponeurosiscompared to modern humans (Morton, 1964; Susman, 1983;Susman et al., 1984; Vereecke et al., 2005). The great apes alsolack a longitudinal arch (Weidenreich, 1923; Elftman and Manter,1935; Morton, 1964; Tuttle, 1970; Vereecke et al., 2003, 2005).

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619 609

Both hard and soft tissue differences between modern humansand the great apes reflect the demands of different positionalbehaviors, and these differences become most recognizablethrough a comparison of terrestrial gaits of modern humans andgreat apes (particularly for Pan, for which most nonhuman data areavailable). During human walking, the heel is the first part of thefoot to contact with the ground. The lateral midfoot touches down,followed by the lateral MT heads and then medial MT heads; thistypical pathway of touchdown is referred to as the medial weighttransfer. As the heel rises, the imposing weight forces the meta-tarsophalangeal joints (MTPJ) into dorsiflexion and the plantaraponeurosis is tightened e an event termed the windlass mecha-nism (Hicks, 1954; Caravaggi et al., 2009). Each phalanx (i.e., thehandle of the windlass) moves onto the dorsum of the respectiveMT head (i.e., the drum of the windlass) and pulls the plantaraponeurosis along with the plantar pad (i.e., cable). When MTPJdorsiflexion increases, the plantar aponeurosis tightens, thelongitudinal arch becomesmore pronounced, and the foot becomesa rigid lever (Hicks, 1954; Bojsen-Møller and Lamoreux, 1979;Erdemir et al., 2004; Caravaggi et al., 2009). As noted in studies ofMT 1, joint compression occurs at the MTPJ during dorsiflexion(Hetherington et al., 1989; Muehleman et al., 1999) and when theMTPJs become extensively dorsiflexed, the collateral ligamentsaround the MTPJ tighten to provide stability, a position known asclose-packing (Susman and Brain, 1988; Susman and de Ruiter,2004). The MTPJs can now serve as a stable fulcrum for pro-pulsion; thus, the human foot can be described as ‘metatarsi-ful-crimating’ (Meldrum, 1991). As suggested by Bojsen-Møller (1978;Bojsen-Møller and Lamoreux, 1979), because the MT 2 is often thelongest of the MTs, forefoot push-off can occur about two axes: thetransverse axis which includes the MT 1 and MT 2 and the obliqueaxis which includes the MT 2 through MT 5. Bojsen-Møller (1978)indicated that the transverse axis is most useful during fast, levelwalking. In addition, the oblique axis is often recruited duringuphill walking, when carrying heavy loads, or at the start of a sprint.The first or second toe is usually the last to break contact with theground at the end of stance phase (Vereecke et al., 2003).

During terrestrial locomotion, foot kinematics of Pan showsmore variation than those of modern humans. The heel and lateralmidfoot of Pan are usually the first parts of the foot to contact thesubstrate, sometimes simultaneously, and these parts are followedby the medial side (Elftman and Manter, 1935; Schmitt and Larson,1995;Wunderlich,1999; Vereecke et al., 2003). However, the halluxsometimes strikes the substrate pointed in the medial direction,and the lateral toes follow in contact (Wunderlich, 1999; Vereeckeet al., 2003). During midstance, the foot can also present with anadducted or abducted first ray, and lateral rays can be eitherplantigrade or curled (Elftman and Manter, 1935; Susman, 1983).As the heel rises, the midfoot does not always break contact withthe ground immediately. This occurrence is called the ‘midtarsalbreak’ or ‘midfoot break’ (DeSilva, 2010), and at this moment, themidfoot serves as a fulcrum (Elftman and Manter, 1935; Susman,1983; Meldrum, 1991; D’Août et al., 2002; Vereecke et al., 2003;DeSilva, 2010)1. As weight is transferred anteriomedially, MTPJs ofPan can extend, but these joints do not experience the close-packing that occurs in the modern human forefoot (Susman et al.,1984; Susman and de Ruiter, 2004). Susman (1983) observed that

1 Originally, the “midtarsal break”was thought to occur when only the transversetarsal joint served as a fulcrum and transmitted weight onto the anterior region ofthe foot (Elftman and Manter, 1935; Nowak et al., 2010), but recent research showsthat more dorsiflexion actually occurs at the cuboid-metatarsal joint (D’Août et al.,2002; Vereecke et al., 2003; DeSilva and MacLatchy, 2008; DeSilva, 2009, 2010)than the calcaneocuboid joint during this event. As a result, DeSilva (2010) hasproposed that the “midtarsal break” be renamed the “midfoot break”.

toeing-off (i.e., during the last part of stance phase when the toesleave the ground) occurs at the second and third digits ofPan troglodytes (chimpanzees) while Wunderlich (1999) noted thatit occurs most often between the first and second digits and lessfrequently at the first or third digit. Vereecke et al. (2003) docu-mented that Pan paniscus (bonobos) can toe-off at the second, thirdor first digit when the foot is not fully pronated (i.e., the hallux isabducted and the lateral toes are curled medially). In sum, themodern human foot contrasts with the foot of Pan by exhibitinga more consistent pattern of leverage during gait e one which ischaracterized by a more stable midfoot, a stereotypical medialweight transfer, a large, adducted MT 1, and MTPJs that forma stable lever at the terminal stage of the stance phase.

Though several researchers have observed that the meta-tarsophalangeal joint dorsiflexion prior to toeing-off in modernhumans is more pronounced than in great apes (e.g., Elftman andManter, 1935; Susman, 1983), the hypothesis that there are statis-tically significant differences between taxa regarding in vivo fore-foot kinematics has not been tested. Tuttle (1970) and Tuttle et al.(1998) measured African ape passive MTPJ (rays 2-5) hyper-dorsi-flexion and noted that if it were not for the long toe length, thesejoints could have a human-like function during gait. Tuttle et al.(1998) further explain that long toes, like those in the great apes,may not be able to support the loading on the forefoot consistentlyduring the terminal stance phase when the MTPJs are highly dor-siflexed. Furthermore, it has been observed that chimpanzees andbonobos can walk with flat, extended toes (Elftman and Manter,1935; Vereecke et al., 2003), but the amount of MTPJ dorsalexcursion coupled with the weight borne on the forefoot atmaximum dorsiflexion has not been quantified. The addition of thisinformationwill more clearly define how amodern humanweight-bearing, metatarsi-fulcrimating foot differs from the great ape footthat lacks this type of forefoot propulsion.

Though there have been previous in vivo studies of humanforefoot kinematics (Bojsen-Møller and Lamoreux, 1979; Kidderet al., 1996; Leardini et al., 1999, 2007; Nawoczenski et al., 1999;MacWilliams et al., 2003; Halstead et al., 2005; Caravaggi et al.,2009), the specific aim of this study is to compare human andbonobo forefoot kinematics. Bonobo forefoot function is of specialinterest because Pan is the closest living relative of humans andserve as an important comparative outgroup for understanding thedescendents of the HomoePan common ancestor (McHenry, 1984;Richmond and Strait, 2000; Vereecke et al., 2003; but see Lovejoyet al., 2009). With the addition of these data, researchers will bein a better position to determine quantitative approaches that aresuitable for interpreting early hominin forefoot function from fossilfoot bones (e.g., Latimer et al., 1982; Lovejoy et al., 2009) andfootprints (e.g., Leakey and Hay, 1979; Bennett et al., 2009).Specifically, this study tests the hypothesis that the great apeforefoot does not dorsiflex in the same manner prior to toe-off asthe human forefoot (Elftman and Manter, 1935; Susman, 1983).Although such a hypothesis has been used as a basis for studies offossil hominin foot function (e.g., Latimer and Lovejoy, 1990;Duncan et al., 1994; Griffin and Richmond, 2010), in vivo forefootkinematics during terrestrial gait need to be empirically tested todetermine whether or not the differences between humans andgreat apes are significant. In addition, new kinematic results will becombined with existing plantar pressure measures in order tofurther highlight what appears to be a unique human metatarsi-fulcrimating mechanism. It was not possible to collect plantarpressure or force plate data in this study, therefore existingbonobo- and human-averaged pressure data profiles were used.The measurement of MTPJ excursion is only one aspect of forefootfunction, and therefore, the combination of angular excursion withplantar pressure of the forefoot will be valuable to studies aiming to

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619610

use skeletal morphology (e.g., trabecular bone architecture) topredict function in fossil hominids.

In order to establish the degree of difference between bonoboand human forefoot mechanics, two predictions were tested. Thefirst prediction is that human first, second, and third meta-tarsophalangeal joints will each show greater amounts of dorsi-flexion prior to toe-off than the corresponding joints of bonobos(whether during terrestrial quadrupedalism or bipedalism).Secondly, because the human forefoot serves as a weight-bearinglever at the terminal stance phase, we predict that the greateraverage MTPJ 1 dorsal excursions are associated with relativelyhigher average hallux plantar pressures during push-off on thehuman foot, while smaller average MTPJ 1 excursions will beassociated with relatively lower average hallux plantar pressureson the bonobo foot.

Sample overview

The human sample comprises two geographic subdivisions andthe bonobo sample is represented by one captive group. The firsthuman sample comes from a previous study (D’Août et al., 2009)and consists of adult male (n ¼ 26) and female (n ¼ 24) inhabitantsof Bangalore, India and nearby areas. This sample can also bedivided into habitually unshod (n ¼ 25) and habitually shod(n ¼ 25) groups (Table 1). The second human sample consists ofhabitually shod adult male (n ¼ 8) and female (n ¼ 8) subjectsresiding in Belgium at the time of the study (Table 1). Permission tosample human subjects was approved by the Ethical Committee ofthe University of Antwerp (#A0401) and the George WashingtonUniversity (IRB #100644).

The Pan paniscus sample comes from a colony housed at theAnimal Park Planckendael in Muizen, Belgium. Only adults wereincluded as subjects (n ¼ 6) (Table 1). The group lives on a 3000 m2

island surrounded by a moat. The island is full of bushes andclimbing structures that mimic a natural habitat and includesupports varying in diameter and inclination (Van Elsacker et al.,1993). Permission to collect bonobo video at the Animal ParkPlanckendael was approved by both the Royal Zoological Society ofAntwerp and the George Washington University (IACUC # 27-11,6).

Methods

We collected three types of video footage. Data capture andprocessing of each type are described below. A description of thestatistical analyses then follows.

Table 1Sample.

Taxon N Age(yrs) *

Mass(kg) *

Height(m)*

Homo sapiens(Bangalore, Índia)Habitually Unshod 25 (16F, 9 M) 46.6 (�13.6) 58.1 (�14.5) 1.59 (�0.09)Habitually Shod 25 (8F, 17 M) 36.9 (�13.6) 59.6 (�7.07) 1.63 (�0.09)

Homo sapiens(Belgium)

16 (8F, 8 M) 30.6 (�8.9) 66.2 (�10.0) 1.74 (�0.11)

Pan paniscus(Belgium)

Subject Age (yrs) SexDJ 12 FHO

e

29 FHE

e

29 FKI

e

29 MZM 12 MRD 17 M

F, female; M, male.* Mean (Standard Deviation).

Humans: India

This sample consists of habitually shod and habitually unshodpeople native to India. Details regarding the sample and method-ology can be found in D’Août et al. (2009). All subjects walkedbarefoot across the viewing area (i.e., lateral to the camera view)and were encouraged to walk with a comfortable velocity. For mostsubjects, three trials were collected per foot. Kinematic data wereonly measured from the right foot when its view was unobstructed(i.e., as individuals moved from right to left across the runway). TheMTPJ angle (defined as the angle formed by the angle of the heel[i.e., the most posterior-plantar point on the heel], the center of theMT head on the medial side, and the distal end of the hallucalproximal phalanx) was digitized in ImageJ software (http://rsbweb.nih.gov/ij/). This angle was measured at midstance and prior to toe-off at maximum dorsiflexion. The midstance frame was chosen asthe moment when the swing leg was almost next to the foot instance, but did not block its view (Fig. 1a). Several frames weredigitized to identify the video frame in which the maximum angleof dorsiflexion occurred prior to toe-off (Fig. 1b). The measurementof MTPJ dorsal excursion is the angle measured at midstance sub-tracted by the angle measured at maximum dorsiflexion. NG digi-tized one trial 10 times (each on separate day) to determineprecision error and standard deviation (SOM Table 1).

Humans and bonobos: Animal Park Planckendael

Filming of the bonobo sample (n ¼ 6) and part of the Belgianhuman sample (n ¼ 8) took place in the Bonobo Playhall at theAnimal Park Planckendael in Muizen, Belgium. Four synchronizedvideo cameras (JVC PAL cameras with Sony Digital-8 recorders,50 Hz) were set up in the playhall in order to collect MTPJ angles in3D (for more details see Griffin, 2009). Trials with a clear view ofthe foot in at least two camera views were saved for digitizationusing the program Motus 7.2.10 (Peak Performance Technologies,Englewood, Colorado). For each trial, locomotor behavior and footposture were recorded (Table 2). MTPJ angle landmark points weresimilar to those chosen for the Bangalore trials, but slightlydifferent due to the use of multiple camera views. In other words,due to the different camera positions in the playhall, the distal endof the proximal phalanx and MT head were digitized at the centerof the joint, not the medial side. In two cases, the foot posture atmidstance was not typical of what would be expected (e.g., the heelwas not in contact with the ground at midstance), and in thesecases another frame exhibiting a more neutral position of the footwas chosen for digitization. Also, it should be noted that in mosttrials (Table 2), the lateral toes exhibit varying degrees of plantarflexion and inversion and are not completely flat or extended on theground. For the bonobo subjects, the “midfoot break”was identifiedduring each cycle.

Angle computation

Over the course of this study, three methods of measuring MTPJangles for the bonobo trials were explored: (1) raw angles gener-ated by the Peak Motus program (Griffin et al., 2008; Griffin, 2009),(2) angles projected onto a reference plane using the Peak Motusprogram (Griffin, 2009) and (3) angles generated using codewritten for the free R Project for Statistical Computing software(“R,” http://www.r-project.org/). As a result of the problemsinherent in the first two measures, the third became the method ofchoice (Griffin, 2009). Essentially, the R program was used to takecoordinate values of landmarks generated in Peak Motus andmeasure the selected angle projected onto a reference planeestablished by the position of the foot. Using the R program, the

Figure 1. The MTPJ 1 angles of habitually shod and unshod individuals from India were digitized in ImageJ. Dorsal joint excursion was defined as the difference of the MTPJ angle at(A) midstance and (B) prior to toe-off when the MTPJ is maximally dorsiflexed. As illustrated in (B), the MTPJ angle is formed by the most posterior-plantar point on the calcaneus,the central point on the medial side of the first metatarsal head, and the distal end of the hallucal proximal phalanx on the medial side. The last two points were marked in black oneach subject prior to filming. Both pictures have been flipped horizontally.

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619 611

reference plane was established by digitized points of two MTheads and the calcaneus (Fig. 2). The line connecting the MT headlandmarks was identified as the normal vector (Fig. 2a), perpen-dicular to the plane of interest (i.e., sagittal plane). This sagittalplane was calculated to pass through the landmark of the calcaneus(Fig. 2a). MTPJ angles were calculated with this reference plane(i.e., landmark points were projected onto this sagittal plane)(Fig. 2b,c,d). For example, in most cases, the MTPJ 2 angle consistedof the angle formed by the calcaneus, MT head segment (MT 1 andMT 2 heads) and the head of second proximal phalanx as projectedonto the reference plane. Different MT head segments were used toform the normal vector (Fig. 2a), depending on forefoot posture and

Table 2Bonobo trial descriptions.

Subject_Trial # Locomotion

DJ_1 knuckle-walking

DJ_2 quadrupedal walk with knuckle-walking hand carrying branches

DJ_3 knuckle-walking

DJ_4 tripedal gallop with baby clinging, and also carrying wood logDJ_5 knuckle-walking with baby clinging

DJ_6 tripedal walk with baby clinging

DJ_7 knuckle-walking, holding branches in one hand

HE_1 knuckle-walking with baby clingingHO_1 knuckle-walkingHO_2 knuckle-walking with baby clingingKI_1 quadrupedal gallopKI_2 tripedal gallop, carrying hay in right handKI_3 quadrupedal gallopRD_1 knuckle-walkingRD_2 knuckle-walkingRD_3 bipedal-walkingRD_4 knuckle-walkingRD_5 knuckle-walkingZM_1 tripedal walk (carrying branches in one hand)ZM_2 tripedal (pushing cardboard box along floor with one hand)ZM_3 tripedal gallopingZM_4 pushing a cardboard box with one hand, no touchdown of

other hand viewedZM_5 knuckle-walkingZM_6 knuckle-walkingZM_7 sliding a cardboard box across the floor with both hands

MT head visibility in the camera. These vectors were, effectively,the axis of rotation of the projected angles because the vertex ofeach angle is always one of the projected MT head landmarks, andthus falls perpendicular to the vector. For one trial, DJ_1, theexcursions of MTPJ 1e3 were measured (Fig. 2). For MTPJ 1, the axiswas defined as the line formed by points representing the MT 1 andMT 2 heads. Formost trials (including all 8 human trials recorded inthe Playhall), the MT 1 and MT 2 segments were used, but for somebonobo trials, other MT head segments were used (SOM Table 2).Regarding the angle measurements of the MTPJ 2 and MTPJ 3 asshown in Fig. 2, the MT 2 andMT 5 head segments were selected. Inthis case, MT 2 and MT 5 were chosen because this particular

Midstance Foot Posture

MT 1 is abducted and lateral digits are slightly flexed; overall, forefoot istending towards plantigradyMT 1 is abducted and lateral digits are slightly flexed; overall, forefoot istending towards plantigradyMT 1 is abducted and lateral digits are slightly flexed; overall, forefootis tending towards plantigradyMT 1 is adducted and plantigrade, lateral digits are flexedMT 1 is abducted and lateral digits are slightly flexed; overall, forefoot istending towards plantigradyMT 1 is abducted and lateral digits are slightly flexed; overall, forefoot istending towards plantigradyMT 1 is abducted and lateral digits are slightly flexed; overall, forefoot istending towards plantigradyMT 1 is abducted and plantigrade, lateral digits are flexedMT 1 is abducted and plantigrade, lateral digits are flexedMT 1 is abducted and plantigrade, lateral digits are flexedMT 1 is abducted and plantigrade; lateral digits show some flexionMT 1 is abducted and plantigrade, lateral digits are flexedMT 1 is abducted and plantigrade; lateral digits show some flexionMT 1 is abducted and plantigrade; lateral digits show a small degree of flexionMT 1 is abducted and plantigrade; lateral digits show some flexionMT 1 is abducted and plantigrade; lateral digits show small degree of flexionMT 1 is abducted and plantigrade; lateral digits show small degree of flexionMT 1 is slightly abducted and is plantigrade, lateral digits are flexedMT 1 is abducted and plantigrade; lateral digits show some flexionMT 1 is abducted and plantigrade, lateral digits are flexedMT 1 is abducted and plantigrade, lateral digits show some flexionMT 1 is abducted and plantigrade, lateral digits show some flexion

MT 1 is abducted and plantigrade; lateral digits show some flexionMT 1 is abducted and plantigrade; lateral digits show a small degree of flexionMT 1 is abducted and plantigrade, lateral digits are flexed

Figure 2. Direction of locomotion and foot posture of bonobo subjects presented a challenge for measuring MTPJ dorsal excursion. Code was written in the program R to providea reference plane that is based on foot posture and ensures that only excursion along the sagittal plane of the forefoot is measured. (A) The plane onto which an MTPJ angle isprojected was formed by two metatarsal head landmarks and the calcaneus landmark. The segment formed by the two metatarsal heads defines the normal vector (MT 1 and MT 2on the left and MT 2 ad MT 5 on the right). The plane of projection is represented by the line formed by the calcaneus and the MT head segment in (A). For this particular trial (DJ_1),raw angles were generated for the MTP1 using the MT 1 and MT 2 head segment (left) while MTPJ 2 and 3 raw angles were generated using the MT 2 and MT 5 head segment (right)(see text for justification). This means that different planes were established for generating the MTPJ 1 and MTPJ 2 excursion values for trial, DJ_1. (B) The code is written to adjustthe forefoot landmarks forming the MTPJ angle to values reflecting a projection along the reference plane (MT 2 and its proximal phalanx are shown only for illustrative purposeswhen the MT 1 and MT 2 head segments are used to generate MTPJ 2 angles for other trials; see SOM Table 3). (C) As shown here for the first ray, after landmarks are projected ontothe plane, the angle is generated. Angles are calculated for all the frames from midstance to toe-off. (D) Here, the MTPJ 1 angle shown represents maximum dorsiflexion prior totoeing-off.

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619612

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619 613

segment appears to be in line with the locations of the MTPJ 2 andMTPJ 3 joints better than the MT 1 and MT 5 segments (see Fig. 2A).The MT 5 head was chosen because it was visible in more than twocameras and was easier to distinguish the MTPJ 5 joint than theMTPJ 3. NG digitized one bonobo trial (ZM_1, listed in Table 2) 10times (each on a separate day) to determine precision error andstandard deviation (SOM Table 3).

The angle representing maximum dorsiflexion was identified asthe smallest angle in the set. Dorsal excursions were calculated thesame way as the Bangalore trials. Of the 25 digitized bonobo trials,six could not be run in the R program to generate the MTPJ 1excursion because an additional MT head was either not easilydistinguished or not visible in two or more camera views. MTPJ 2excursions could bemeasured for eight trials. In the other cases, thesecond ray was substantially flexed or not visible in more thanone camera view. Only one trial was suitable for digitization of theMTPJ 3.

Velocity could not be directly calculated for the trials collected atPlanckendael because of restrictions to camera views. Contact timewas measured for the purpose of determining the timing ofmaximum dorsiflexion during the stance phase. Previous studieshave shown an inverse correlation between contact time andvelocity for humans (Vilensky and Gehlsen, 1984) and great apes(Demes et al., 1994; Vereecke et al., 2004).

Humans: Brussels

Human data were also collected using a VICON 612 system(http://www.vicon.com/spots/ESMAC2009.html) at the Universityof Brussels Biomechanics Lab (for more details see Hagman et al.,2008; Griffin, 2009). The sample consisted of 10 subjects, withtwo of the subjects having also been subjects for the video collectedat Planckendael. Subjects were asked to walk and jog ata comfortable pace (for more details, see Griffin, 2009). Fourteenmillimeter infrared markers were attached to the right side of thebody (Fig. 3). Onemarker was placed on themedial side of the headof the first MTand another on the lateral side of the head of the fifthMT to create a segment. Then, a marker was placed on the dorsalaspects of each of the distal ends of first three proximal phalanges.This setup allowed for the measurement of the MTPJ 1, 2, and 3angles using the MT 1 and MT 5 segment as the vertex in each case.For one subject with exceptionally short digits and overall foot size,two trials were necessary to record kinematics and a slightlydifferent marker system was used (see Griffin, 2009). Like thePlanckendael trials, MTPJ angles of all the VICON trials weregenerated using the R program and using the calcaneus and the twoMT head coordinates to form the reference plane. Summarystatistics for each sample’s raw angles (i.e., midstance angle andangle at greatest MTPJ dorsiflexion) can be found in SOM Table 4.

Analyses

Statistical summaries for joint excursion values were generatedin Statistica 7 (Statsoft, Tulsa, OK). A comparison of MTPJ 1 excur-sions was first made between the two Bangalore groups in order totest whether a lifetime of wearing shoes may have had an effect onjoint range of motion during barefoot walking. Velocity andexcursion ranges were averaged when two or three trials had beenrecorded for a subject. To test for a difference in the habitually shodand habitually unshod groups, values of MTPJ excursion wereregressed against values of velocity. Least squares regression lineswere generated for each group and compared using an ANCOVA.

Before interspecific differences in metatarsophalangeal jointexcursion were explored, first metatarsophalangeal joint excursionvalues and contact time values were each averaged per human

subject when subjects participated in more than one trial. For theVICON video, subjects participated in both walking and joggingtrials and therefore, walking and jogging trials were averagedseparately. As a result of the small sample size available, bonobotrials were not averaged per subject. The human and bonoboexcursions were compared using the nonparametric Watson’s U2

test for angular data (Zar, 1999).Finally, the timing of both maximum MTPJ 1 dorsiflexion and

the “midfoot break” during the stance phase were expressed aspercentages of contact time (i.e., when during stance phase, the“midfoot break” and MTPJ excursion occur). This permitted inclu-sion with Vereecke et al.’s (2003: Fig. 4) averaged plantar pressureprofiles. Here, plantar pressure is assumed to be an indication ofloading. Timings of maximumMTPJ 1 dorsiflexion (i.e., percentagesof the stance phase) were averaged for the quadrupedal walkingtrials of bonobos and separately for the bipedal walking trials of thehumans. The occurrence of the “midfoot break” was also averagedfor bonobo quadrupedal walking and bipedal walking trials. Foreach locomotion type, event timings were first averaged by subjectbefore the taxonomic average of either MTPJ 1 maximum dorsi-flexion or “midfoot break” was calculated. For the human sample,only the Bangalore and Planckendael videos were used becausecontact time could be estimated with greater accuracy than VICONvideo. Only one bonobo bipedal trial was available, so its valueswere used. Timings of either MTPJ 2 or MTPJ 3 maximum dorsi-flexion relative to plantar pressure were not considered becausethe pressure data were not available for each of these individualrays.

Results

As predicted, human metatarsophalangeal joints dorsiflex morethan those of bonobos, and coupled with the plantar pressure, thetiming of maximum dorsiflexion of the first metatarsophalangealjoint highlights the key role of the human forefoot in propulsion.Summary statistics for human and bonobo samples are reported inTables 3e5. Although the habitually shod human group has bothhigher mean MTPJ excursion and velocity values than the habitu-ally unshod group (Table 3), the comparison of the regressionsgenerated for each group showed neither slope differences(p ¼ 0.47) nor y-intercept differences (p ¼ 0.12) (Fig. 4). This resultjustified combining the habitually shod and habitually unshod datafor the interspecific comparisons.

The ranges of the human forefoot excursions are: MTPJ 1,24�e53�; MTPJ 2, 25�e47� andMTPJ 3,16�e42�. The bonoboMTPJ 1values range from 10�e38� and the MTPJ 2 values range from14�e34�. The one bonoboMTPJ 3 dorsal excursionwas measured as25 degrees, and is well within the excursion range for the humanMTPJ 3. Though there is overlap between the taxa regarding rangesof excursions, as expected, human MTPJ 1 and MTPJ 2 excursionsare significantly larger than those of bonobo MTPJ 1 and 2(p < 0.001 in both cases) (Fig. 5).

The timing of forefoot kinematics and existing plantar pressuredata also distinguish the two taxa (Fig. 6). For example, the timingof maximum MTPJ 1 dorsiflexion occurs at an average of 87%(�2.1%) of contact time for a quadrupedal bonobo, 93% of contacttime for the one bipedal bonobo, and an average of 96% (�2.5) ofcontact time for the pooled human trials. Timing of maximumdorsiflexion in the human sample takes place just before peakpressure underneath the first ray, which occurs at 98% of contacttime. For the quadrupedal bonobo, peak pressure under the first rayoccurs at 84% of contact time, which is also around the time ofmaximumMTPJ 1 dorsiflexion. In contrast to the human trials, bothquadrupedal (Fig. 6A) and bipedal trials (Fig. 6B) indicate thatmaximum dorsiflexion is not associated with a substantial increase

Figure 3. Ten human subjects were recruited at the University of Brussels Biomechanics Lab in order to measure MTPJ kinematics during walking and jogging. (A) Fourteenmillimeter infrared markers were attached to the right side of the body on the greater trochanter, the lateral and medial femoral condyles, the lateral and medial malleoli, thecalcaneal tuberosity, the navicular tuberosity, the heads of the first and fifth metatarsals and the heads of the first, second and third proximal phalanges. (B) Close-up of the footmarkers (marker for calcaneus is not in view).

Table 3MTPJ 1 summary statistics.

Sample N Dorsal Excursion Velocity (m/s) Contact time (s)

Mean & SD Range Mean & SD Mean & SD Range

Homo sapiensHabitually Unshod (India) 25 37�6.2 24e51 0.96�0.25 0.87�0.17 0.61e1.3Habitually ShodIndia 25 41�5.8 34e53 1.1�0.23 0.77�0.13 0.61e1.3Planckendael 8 40�6.2 29e47 n/a 0.73�0.059 0.62e0.78VICONWalking 10 43�4.9 35e48 n/a 0.70�0.095 0.63e0.93Jogging 10 36�4.8 30e43 n/a 0.29�0.051 0.21e0.38

Pan paniscusQuadrupedalism 5 19�5.5 10e24 n/a 0.86�0.064 0.77e0.94Tripedalism 1 13 n/a n/a 0.89 n/aQuadrupedal/Tripedal Gallop 3 26�10 18e38 n/a 0.34�0.020 0.32e0.36Bipedalism 1 18 n/a n/a 0.88 n/a

SD ¼ Standard Deviation.n/a ¼ not available.Two human individuals participated as subjects in both the Animal Park Planckendael and VICON setups. For the statistical analysis, the excursions of the walking trials fromeach setup were averaged together. They are not averaged together in the table above.

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619614

Table 4MTPJ 2 Summary Statistics.

Sample N Dorsal Excursion Contact time(s)

Mean & SD Range Mean & SD Range

Homo sapiensPlanckendael 8 39�6.4 30e47 0.73�0.059 0.62e0.78VICONWalking 10 34�4.3 25e38 0.66�0.14 0.41e0.93Jogging 10 33�4.2 25e41 0.29�0.050 0.21e0.38

Pan paniscusQuadrupedalism 3 24�10 14e34 0.77�0.095 0.66e0.83Tripedalism 1 19 n/a 0.97�0.21 0.82e1.1Quadrupedal Gallop 1 22 n/a 0.28 n/a

SD ¼ Standard Deviation.n/a ¼ not available.Two human individuals participated as subjects in both the Animal Park Planck-endael and VICON setups. For the statistical analysis, the excursions of the walkingtrials from each setup were averaged together. They are not averaged together inthe table above.

Figure 4. A comparison of the slope generated by the habitually unshod Indian group(stippled line, y ¼ 32.7 þ 4.6x) and the slope of the habitually shod Indian group (solidline, y ¼ 30.1 þ 9.7x) indicates that relative to velocity, the two groups do not differ inMTPJ 1 dorsal excursion prior to toe-off. Habitually unshod human values are repre-sented by open circles; habitually shod human values are represented by blackdiamonds.

1.41.21.00.80.60.40.20.00

10

20

30

40

50

60JPTM .sv emiT tcatnoC noisrucxE lasroD 1

lasroDnoisrucxE)seerged(

)s( emiT tcatnoC

4_JD

1_IK

3_MZ2_IK

3_IK

5_MZ

6_MZ 1_JD1_OH

4_MZ 2_JD

3_JD3_DR1_EH

2_MZ1_DR

7_MZ2_DR

1_MZ

4_JD

1_IK

3_MZ2_IK

3_IK

5_MZ

6_MZ 1_JD1_OH

4_MZ 2_JD

3_JD3_DR1_EH

2_MZ1_DR

7_MZ2_DR

1_MZ

4_JD

1_IK

3_MZ2_IK

3_IK

5_MZ

6_MZ 1_JD1_OH

4_MZ 2_JD

3_JD3_DR1_EH

2_MZ1_DR

7_MZ2_DR

1_MZ

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619 615

in pressure under the first ray (relative to the lateral rays) leading toa spike as shown in the human trial profile. Therefore, the spike ofpressure distinguishes the human first ray from the rest of the footand clearly shows its primary role in weight-bearing just prior totoeing-off (Fig. 6C). MTPJ 1 maximum dorsiflexion in the bipedalbonobo occurs when pressure in the first ray is already decliningand the pressure increases in the lateral rays and lateral midfoot.While the quadrupedal bonobo profile shows a human-like patternwith maximum dorsiflexion taking place when the first digit bearsmore pressure than any other region of the foot, there is nota substantial difference between the pressure of the first ray andthe lateral rays present in the human profile.

The “midfoot break” was observed at around 64% (�11.7%) ofcontact time for the quadrupedal bonobo trials and 70% of contacttime for the bipedal bonobo (Fig. 6A,B). At the “midfoot break”during the stance phase, pressure is present under the lateralmidfoot, and as more clearly shown in the quadrupedal bonoboprofile, this corresponds with the role of the midfoot as a weight-bearing site during stance phase of Pan. Neither human subjects inthe current study sample nor those in Vereecke et al.’s (2003) studyexhibited a “midfoot break.” In the latter study the lack of a “mid-foot break”was reflected in the minimal amount of weight supportby the lateral midfoot during stance phase.

Discussion

Habitual bipedalism is a key hominin adaptation. Traditionally,the external morphology of the lower limb has been at the center ofdebates over whether or not early hominin gait was modernhuman-like and whether or not climbing was an important part oflocomotor behavior. However, relying onmorphological differencesalone has proven to be inconclusive so far, especially regarding thefunctional implications of forefoot morphology among early

Table 5MTPJ 3 summary statistics.

Sample N Dorsal Excursion Contact time (s)

Mean & SD Range Mean & SD Range

Homo sapiens (VICON)Walking 10 31�7.5 16e42 0.70�0.095 0.63e0.93Jogging 10 30�6.2 17e39 0.29�0.051 0.21e0.38

Pan paniscusQuadrupedalism 1 25 n/a 0.80 n/a

SD ¼ Standard Deviation.n/a ¼ not available.

hominins. This obstacle is further compounded by the lack of dataon forefoot function and structure in living great apes and humans.In light of these problems, the main goal of this study was to testwhether or not human MTPJs show greater dorsal excursions priorto toe-off than theMTPJs of bonobos during terrestrial gait. Keepingsuch limitations as the small and variable (in terms of gait type andfoot posture; see Table 2) bonobo sample in mind, the observationsfor theMTPJ 1 and 2 support the conclusion that, during gait, dorsalexcursion of human forefoot generally exceeds that of the bonobo

:)aidnI ,erolagnaB( snamuH

:)muigleB ,leadnekcnalP( snamuH

:)muigleB ,slessurB( snamuH dohS yllautibaH

:)muigleB ,leadnekcnalP(sobonoB

+

Figure 5. The human sample is composed of the Indian habitually unshod and shodsample (open circles), the Planckendael human sample (closed squares), the Brusselssample (crosses), and the average values of the individuals that participated in both thePlanckendael and Brussels walking trials (closed circles). Each bonobo value is repre-sented by a closed triangle followed by a code indicating the individual (see Tables 1and 2), trial number (see Table 2). Note the values from the Brussels sample formtwo distinct clusters because walking and jogging trials of each subject are repre-sented. It should also be noted that some bonobos achieve values of excursioncomparable to those of humans.

Figure 6. The timing of maximum MTPJ 1 dorsiflexion (solid arrow) during the stance phase of the (A) averaged quadrupedal bonobo trials, (B) bipedal bonobo trial, and(C) averaged walking human trials were matched with the corresponding averaged plantar pressure profiles (of the hallux [first metatarsal and first digit]), courtesy of Vereeckeet al. (2003). Maximum dorsiflexion (solid arrows) occurs at 87% (�2.1%) of stance phase for the averaged quadrupedal bonobo, 93% for the one bipedal bonobo, and 96% (�2.5)contact time for the averaged human walking trials. Note that maximum MTPJ 1 dorsiflexion in humans occurs just before the first ray experiences peak pressure (at 98% contacttime). The average occurrence of the ‘midfoot break’ (broken arrow) is 64% (�11.7%) of contact time for the quadrupedal bonobos and 70% of contact time for the one bipedalbonobo. On a separate note, the plantar pressure profile of the bipedal bonobo indicates that plantar pressure of the heel is present throughout the stance phase. As indicated in theplantar pressure profile, weight is not only transferred to the anterior part of the foot, but plantar pressure in the hindfoot region also shows a slight increase as well. This suggeststhat this averaged plantar pressure profile includes individual bonobos who did not exhibit a ‘midfoot break’. When a midtarsal break occurs, pressure under the heel region shouldbe absent. Between 64% and 70% contact time, the human lateral midfoot exhibits minimal plantar pressure. H, heel; V1, medial midfoot; V2, lateral midfoot; M, metatarsal heads2e5; T, lateral toes 2e5; Hallux, first metatarsal head and first digit.

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619616

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619 617

forefoot. Additional bonobo trials need to be collected beforetaxonomic differences in MTPJ 3 excursions can be tested.

The results support previous observations that the forefoot ofPan does not dorsiflex in the same manner as the human forefoot(Elftman and Manter, 1935; Susman, 1983). The results also showsome previously unrecognized levels of overlap between the taxa inranges of excursion for bothMTPJ 1 and 2, and in future studies, thisoverlapmay become accentuatedwith the addition ofmore bonobosubjects and trials. Also, the excursion of the onebonoboMTPJ 3 fallswithin the observed human range. These areas of overlap may beinfluenced by some methodological limitations. For example, it ispossible that the greater joint mobility in the bonobo midfoot(i.e., dorsiflexion as the heel rises during the stance phase)mayhavecontributed to a larger value ofMTPJ dorsal excursion thanwould bemeasured if the base of the MT was used as a marker instead.Because markers could not be placed on the bonobo subjects, thecalcaneus was chosen as a proxy for the MT base in all of thesubjects. The calcaneus has beenused in other studies to reconstructMTPJ excursion (e.g., Nawoczenski et al., 1999), but may introduceerrors in comparisons between taxa with very different foot struc-tures. Another limitation is the use of the MT head segment as thevertex, which further prevents the direct measurement of MTPJmotion, especially in the case of the bonobos when the hallux isdivergent from the second ray. Ideally, both taxonomic sampleswould be measured with a VICON system and markers wouldbe placed at the locations defining the metatarsophalangeal joint(i.e., head of the proximal phalanx, MT head, and MT base).

Though it is important to acknowledge the overlap in MTPJexcursion values between the two taxa, this study supports andadds to previous research. Coupling the present study’s kinematicdata with Vereecke et al.’s (2003) pressure data clearly distin-guishes human forefoot function from that of the bonobo. Not onlyis human maximum dorsiflexion on average greater than that ofthe bonobo, the pressure at this occurrence is relatively greater(i.e., hallux vs. lateral forefoot pressure) than under the bonobo firstray during bipedal or quadrupedal walking (Fig. 6). Thus, themedialside of the human forefoot contrasts with the bonobo medialforefoot by showing a larger contribution to propulsion during gait.

When the occurrence of the “midfoot break” in the currentstudy’s trials is included with Vereecke et al.’s (2003) pressure data,the role of the lateral midfoot during weight-bearing in bonobos ishighlighted (Fig. 6) (Elftman and Manter, 1935). However, thefulcrum action of the lateral midfoot is not as clear. For example,regarding the single bipedal bonobo trial, the occurrence of the“midfoot break” does not match the plantar pressure profilereported by Vereecke et al. (2003). In the average plantar pressureprofile, the pressure under the heel is present through most ofthe stance phase, indicating that in this profile that at least some ofthe subjects did not exhibit a “midfoot break.” Also, when theoccurrence of the “midfoot break” is matched on the quadrupedalplantar pressure profile, it shows that plantar pressure in the lateralmidfoot is decreasing as weight is being transferred to the anteriorregions of the foot. This is not comparable to the human pattern inthe forefoot, in which maximum MTPJ 1 dorsiflexion occurs whenpressure underneath the first ray is increasing.

An overarching goal of this study is to increase our currentknowledge about forefoot functional differences in extant hominidsfor the purpose of making inferences about early hominin posturaland locomotor behavior. Several in vivo studies have shown thattrabecular struts and plates can respond to the direction andmagnitude of local loading byaligning themselves in the directionofstress and increasing in density (Lanyon,1974; Goldstein et al.,1991;Pontzer et al., 2006; van der Meulen et al., 2006). In addition, high-resolution X-ray computed tomography has shown significantdifferences in trabecular bone properties betweenprimates that use

habitually different locomotor behaviors (Fajardo andMüller, 2001;MacLatchy and Müller, 2002; Ryan and Ketcham, 2002a,b, 2005).The kinematic data presented in this study support results from theanalysis of trabecular architecture in the extant hominidMTPJ 1 and2 (Griffin et al., 2010). Quantification of the trabecular bone of eitherthe MT 1 or MT 2 head shows that humans can be distinguishedfrom the great apes based on a more organized trabecular struttingpattern, especially in the dorsal region. The more stereotypicalpattern of strutting and reinforcement of bone in the dorsal area ofthe human joint correspond with its role as a primary lever priorto toe-off. The less organized pattern of the great ape trabeculaeemphasizes that the MTPJs need to accommodate stresses inseveral directions as a result arboreal foot use for grasping and invariable postures during terrestrial gait (Table 2). Combining thecurrent experimental results with morphological observationssuggests that trabecular bone structure in hominin forefoot fossilsmay provide a means of testing hypotheses about the origin andevolution of the human metatarsi-fulcrimating mechanism.

Together, the results of the kinematic study and more recentstudies such as the measurement of dorsal canting in proximalphalanges (Griffin and Richmond, 2010) show that human MTPJsusually exhibit more excursion, which is reflected in the previousand current findings that human proximal phalanges have moredorsally canted basal joints than those of Pan. However, it isimportant to note that there is overlap in ranges of the two taxaregarding results in both kinematic and morphological analyses.These results affirm Tuttle’s (1970) and Tuttle et al.’s (1998) findingthat passive dorsiflexion of the African ape MTPJs is comparable tothat of humans. As noted in the current and previous studies, thisMTPJ dorsiflexion is not habitual because bonobo and chimpanzeesubjects oftenwalk with curled toes (i.e., flexed phalanges) and thisposture is most likely used to avoid high bending stresses associ-ated resulting from having long toes (Tuttle, 1970; Preuschoft,1970). Both the overlap in dorsal canting, as well as the in vivoand in vitro kinematic data, suggest that humans are not unique inthe ability to dorsiflex their MTPJs quite substantially. Also,according to Rein and Harrison (2007), some species of Papio areindistinguishable from humans in the degree of dorsal canting.Thus, simply observing or quantifying dorsal canting of a singlephalanx should not be used as the only diagnostic character ofhuman-like forefoot function. However, it can be used in combi-nation with other traits, specifically those that will provide clues towhether or not the forefoot was being used as a primary weight-bearing lever. This includes a specific pattern of trabecular boneorientation in the dorsal region of the MT head (Griffin et al., 2010).

Conclusions

This study is the first to provide quantitative measurements ofboth human and bonobo forefoot kinematics during locomotion.The results support previous qualitative observations that great apeMTPJs do not usually dorsiflex in the same manner as humans priorto toe-off. In addition, relating the occurrence of a “midfoot break”to the pressure profile of the lateral midfoot of Pan emphasizes theuse of this region for weight transmission. However, the foot of Pandoes not exhibit a notable push-off mechanism like the humanfoot; the end of stance phase in Pan might be better described as“lift-off” than “push-off.” The results here show that the role of thehuman forefoot as a key site of leverage and weight transmission isunique compared to our closest living relatives.

Acknowledgments

We wish to thank the Royal Zoological Society of Antwerp andthe Animal Park Planckendael staff for permitting KD and NG to

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619618

videotape the bonobos. We are grateful for the zookeepers’ coop-eration and patience, and thankful to all of our subjects. Specialthanks is given to Jan Scholliers for constructing the calibrationcube, Dave Frans for deinterlacing the Bangalore video, and Dr. BartVan Gheluwe for allowing KD and NG to use the VICON system inhis Biomechanics Lab. NG is grateful to Adrienne and Jan Van DenBaere for their exceptional hospitality during her stay in Muizen,Gabriel Mallozzi for his assistance with Peak Motus, Can Kirmizi-bayrak and Erin Marie Williams for their help with the VICONprogram, J. Tyler Faith for his statistical help and Pascal W. Prinz formodeling his own hindlimb. We greatly appreciate Dr. Evie Ver-eecke for making the bonobo and human plantar pressure dataavailable. We are grateful to Daniel Schmitt, Masato Nakatsukasa,Bernard Wood, and Tim Ryan for their comments on drafts of thismanuscript. We are grateful to Steven Leigh, and the anonymousAssociate Editor and reviewers for their editorial suggestions.

This study has been supported by the L.S.B. Leakey Foundation,NSF BCS-0726124, Sigma Xi Grant-in-Aid of Research, Lewis CotlowResearch Fund, and through PA and KD, the FWO-Flanders GrantG.0125.05, and BOF-UA. NG would also like to acknowledgethe GWU Selective Excellence Initiative, NSF IGERT grants DGE-9987590 and DGE-0801634 as well as Duke University.

Appendix. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.jhevol.2010.07.017.

References

Alexander, R.M., 2004. Bipedal animals, and their differences from humans. J. Anat.204, 321e330.

Bennett, M.R., Harris, J.W., Richmond, B.G., Braun, D.R., Mbua, E., Kiura, P., Olago, D.,Kibunjia, M., Omuombo, C., Behrensmeyer, A.K., Huddart, D., Gonzalez, S., 2009.Early hominin foot morphology based on 1.5-million-year-old footprints fromIleret, Kenya. Science 323, 1197e1201.

Bojsen-Møller, F., 1978. The Human Foot, a Two Speed Construction. In: Asmus-sen, E., Jørgensen, K. (Eds.), International Series of Biomechanics. UniversityPark Press, Baltimore, pp. 261e266.

Bojsen-Møller, F., 1979. Calcaneocuboid joint and stability of the longitudinal arch ofthe foot at high and low gear push off. J. Anat. 129, 165e176.

Bojsen-Møller, F., Lamoreux, L., 1979. Significance of free-dorsiflexion of the toes inwalking. Acta Orthop. Scand. 50, 471e479.

Caravaggi, P., Pataky, T., Goulermas, J.Y., Savage, R., Crompton, R., 2009. A dynamicmodel of the windlass mechanism of the foot: evidence for early stance phasepreloading of the plantar aponeurosis. J. Exp. Biol. 212, 2491e2499.

Crompton, R.H., Vereecke, E.E., Thorpe, S.K., 2008. Locomotion and posture from thecommon hominoid ancestor to fully modern hominins, with special referenceto the last common panin/hominin ancestor. J. Anat. 212, 501e543.

D’Août, K., Aerts, P., De Clercq, D., De Meester, K., Van Elsacker, L., 2002. Segmentand joint angles of hind limb during bipedal and quadrupedal walking of thebonobo (Pan paniscus). Am. J. Phys. Anthropol. 119, 37e51.

D’Août, K., Pataky, T., De Clercq, D., Aerts, P., 2009. The effects of habitualfootwear use: foot shape and function in native barefoot walkers. FootwearSci. 1, 81e94.

Demes, B., Larson, S.G., Stern Jr., J.T., Jungers, W.L., Biknevicius, A.R., Schmitt, D.,1994. The kinetics of primate quadrupedalism: ‘hindlimb drive’ reconsidered.J. Hum. Evol. 26, 353e374.

DeSilva, J.M., 2009. Vertical Climbing Adaptations in the Anthropoid Ankle andMidfoot: Implications for Locomotion in Miocene Catarrhines and PlioePleis-tocene Hominins. Ph.D dissertation, University of Michigan.

DeSilva, J.M., 2010. Revisiting the “midtarsal break”. Am. J. Phys. Anthropol. 141,245e258.

DeSilva, J.M., MacLatchy, L.M., 2008. Revisiting the midtarsal break. Am. J. Phys.Anthropol. 46, 89.

Duncan, A.S., Kappelman, J., Shapiro, L.J., 1994. Metatarsophalangeal joint function andpositional behavior in Australopithecus afarensis. Am. J. Phys. Anthropol. 93, 67e81.

Elftman, H., Manter, J., 1935. Chimpanzee and human feet in bipedal walking.Am. J. Phys. Anthropol. 20, 69e79.

Erdemir, A., Hamel, A.J., Fauth, A.R., Piazza, S.J., Sharkey, N.A., 2004. Dynamicloading of the plantar aponeurosis in walking. J. Bone Jt. Surg. Am. 86 (A),546e552.

Fajardo, R.J., Müller, R., 2001. Three-dimensional analysis of nonhuman primatetrabecular architecture using micro-computed tomography. Am. J. Phys.Anthropol. 115, 327e336.

Goldstein, S.A., Matthews, L.S., Kuhn, J.L., Hollister, S.J., 1991. Trabecular boneremodeling: an experimental model. J. Biomech. 24 (Suppl. 1), 135e150.

Griffin, N.L., 2009. Comparative Forefoot Kinematics and Bone Architecture inExtant Hominids. Ph.D dissertation, George Washington University.

Griffin, N.L., D’Août, K., Aerts, P., Richmond, B.G., 2008. Comparative in vivo forefootkinematics in extant hominids. Am. J. Phys. Anthropol. 46, 108.

Griffin, N.L., D’Aout, K., Ryan, T.M., Richmond, B.G., Ketcham, R.A., Postnov, A., 2010.Comparative forefoot trabecular bone architecture in extant hominids.J. Hum. Evol.

Griffin, N.L., Richmond, B.G., 2010. Joint orientation and function in great ape andhuman proximal pedal phalanges. Am. J. Phys. Anthropol. 141, 116e123.

Hagman, F., van de Ven, F., Van Gheluwe, B., van Eijndhoven, S., 2008. Setting upgait analysis measurement devices for simultaneously measuring force, pres-sure, and motion. In: D’Août, K., Lescrenier, B., Van Gheluwe, B., De Clercq, D.(Eds.), Advances in Plantar Pressure Measurements in Clinical and ScientificResearch. Shaker Publishing BV, pp. 26e43.

Halstead, J., Turner, D.E., Redmond, A.C., 2005. The relationship between halluxdorsiflexion and ankle joint complex frontal plane kinematics: a preliminarystudy. Clin. Biomech. 20, 526e531 (Bristol, Avon).

Hetherington, V.J., Carnett, J., Patterson, B.A., 1989. Motion of the first meta-tarsophalangeal joint. J. Foot Surg. 28, 13e19.

Hicks, J.H., 1954. The mechanics of the foot. II. The plantar aponeurosis and the arch.J. Anat. 88, 25e30.

Ker, R.F., Bennett, M.B., Bibby, S.R., Kester, R.C., Alexander, R.M., 1987. The spring inthe arch of the human foot. Nature 325, 147e149.

Kidder, S.M., Abuzzahab Jr, F.S., Harris, G.F., Johnson, J.E., 1996. A system for theanalysis of foot and ankle kinematics during gait. IEEE Trans. Rehabil. Eng. 4,25e32.

Lanyon, L.E., 1974. Experimental support for the trajectorial theory of bone struc-ture. J. Bone Jt. Surg. Br. 56, 160e166.

Latimer, B.M., Lovejoy, C.O., Johanson, D.C., Coppens, Y., 1982. Hominid tarsal,metatarsal, and phalangeal bones recovered from the Hadar Formation:1974e1977 collections. Am. J. Phys. Anthropol. 57, 701e719.

Latimer, B., Lovejoy, C.O., 1990. Metatarsophalangeal joints of Australopithecusafarensis. Am. J. Phys. Anthropol. 83, 13e23.

Leakey, M.D., Hay, R.L., 1979. Pliocene footprints in the Laetolil Beds at Laetoli,northern Tanzania. Nature, 317e323.

Leardini, A., Benedetti, M.G., Catani, F., Simoncini, L., Giannini, S., 1999. Ananatomically based protocol for the description of foot segment kinematicsduring gait. Clin. Biomech. 14, 528e536 (Bristol, Avon).

Leardini, A., Sawacha, Z., Paolini, G., Ingrosso, S., Nativo, R., Benedetti, M.G., 2007.A new anatomically based protocol for gait analysis in children. Gait Posture 26,560e571.

Lovejoy, C.O., Latimer, B., Suwa, G., Asfaw, B., White, T.D., 2009. Combiningprehension and propulsion: the foot of Ardipithecus ramidus. Science 326 (72),e1ee78.

MacLatchy, L., Müller, R., 2002. A comparison of the femoral head and necktrabecular architecture of Galago and Perodicticus using micro-computedtomography (microCT). J. Hum. Evol. 43, 89e105.

MacWilliams, B.A., Cowley, M., Nicholson, D.E., 2003. Foot kinematics and kineticsduring adolescent gait. Gait Posture 17, 214e224.

McHenry, H., 1984. The common ancestor: a study of the postcranium of Panpaniscus, Australopithecus, and other hominoids. In: Susman, R.L. (Ed.),The Pygmy Chimpanzee: Evolutionary Biology and Behavior. Plenum Press,New York, pp. 201e230.

Meldrum, D.J., 1991. Kinematics of the cercopithecine foot on arboreal and terres-trial substrates with implications for the interpretation of hominid terrestrialadaptations. Am. J. Phys. Anthropol. 84, 273e289.

van der Meulen, M.C., Morgan, T.G., Yang, X., Baldini, T.H., Myers, E.R., Wright, T.M.,Bostrom, M.P., 2006. Cancellous bone adaptation to in vivo loading in a rabbitmodel. Bone 38, 871e877.

Morton, D.J., 1964. The Human Foot: Its Evolution, Physiology and FunctionalDisorders. Hafner Publishing Company, Inc., New York.

Muehleman, C., Bareither, D., Manion, B.L., 1999. A densitometric analysis of thehuman first metatarsal bone. J. Anat. 195 (Pt 2), 191e197.

Nawoczenski, D.A., Baumhauer, J.F., Umberger, B.R., 1999. Relationship betweenclinical measurements and motion of the first metatarsophalangeal joint duringgait. J. Bone Jt. Surg. Am. 81, 370e376.

Nowak, M.G., Carlson, K.J., Patel, B.A., 2010. Apparent density of the primate cal-caneo-cuboid joint and its association with locomotor mode, foot posture, andthe midtarsal break. AJPA 142, 180e193.

Pontzer, H., Lieberman, D.E., Momin, E., Devlin, M.J., Polk, J.D., Hallgrimsson, B.,Cooper, D.M., 2006. Trabecular bone in the bird knee responds with highsensitivity to changes in load orientation. J. Exp. Biol. 209, 57e65.

Preuschoft, H., 1970. Functional anatomy of the lower extremity. In: Bourne, G.H.(Ed.), The Chimpanzee, vol. 3. Karger, Basel, pp. 221e294.

Rein, T.R., Harrison, T., 2007. Quantifying the angle of orientation of the meta-tarsophalangeal joint surface of proximal phalanges in extant primates.Am. J. Phys. Anthropol. 44, 197.

Richmond, B.G., Strait, D.S., 2000. Evidence that humans evolved from a knuckle-walking ancestor. Nature 404, 382e385.

N.L. Griffin et al. / Journal of Human Evolution 59 (2010) 608e619 619

Rolian, C., Lieberman, D.E., Hamill, J., Scott, J.W., Werbel, W., 2009. Walking, runningand the evolution of short toes in humans. J. Exp. Biol. 212, 713e721.

Ryan, T.M., Ketcham, R.A., 2002a. The three-dimensional structure of trabecularbone in the femoral head of strepsirrhine primates. J. Hum. Evol. 43, 1e26.

Ryan, T.M., Ketcham, R.A., 2002b. Femoral head trabecular bone structure in twoomomyid primates. J. Hum. Evol. 43, 241e263.

Ryan, T.M., Ketcham, R.A., 2005. Angular orientation of trabecular bone in thefemoral head and its relationship to hip joint loads in leaping primates.J. Morphol. 265, 249e263.

Schmitt, D., Larson, S.G., 1995. Heel contact as a function of substrate type and speedin primates. Am. J. Phys. Anthropol. 96, 39e50.

Sockol, M.D., Raichlen, D.A., Pontzer, H., 2007. Chimpanzee locomotor energetics andthe origin of human bipedalism. Proc. Natl. Acad. Sci. USA 104, 12265e12269.

Susman, R.L., 1983. Evolution of the human foot: evidence from PlioePleistocenehominids. Foot Ankle 3, 365e376.

Susman, R., Brain, T., 1988. New first metatarsal (SKX 5017) from Swartkrans andthe gait of Paranthropus robustus. Am. J. Phys. Anthropol. 77, 7e15.

Susman, R.L., Stern Jr., J.T., Jungers, W.L., 1984. Arboreality and bipedality in theHadar hominids. Folia Primatol. 43, 113e156.

Susman, R.L., de Ruiter, D.J., 2004. New hominin first metatarsal (SK 1813) fromSwartkrans. J. Hum. Evol. 47, 171e181.

Tuttle, R.H., 1970. Postural, propulsive, and prehensile capabilities in the cheiridia ofchimpanzees and other great apes. In: Bourne, G. (Ed.), The Chimpanzee.Karger, Basel, New York, pp. 167e253.

Tuttle, R.H., Hallgrimsson, B., Stein, T., 1998. Heel, squat, stand, stride: function andevolution of hominoid feet. In: Strasser, E., Fleagle, J., Rosenberger, A.,McHenry, H. (Eds.), Primate Locomotion: Recent Advances. Plenum Press, NewYork, pp. 435e448.

Van Elsacker, L., Claes, G., Melens, W., Struyf, K., Vervaecke, H., Walraven, V., 1993.New outdoor exhibit for a bonobo group at Planckendael design and intro-duction procedures. In: Daman, F.J. (Ed.), Bonobo Tidings, Jubilee Volume on theOccasion of the 150th Anniversary of the Royal Zoological Society of Antwerp.Royal Zoological Society of Antwerp, Antwerp, pp. 35e47.

Vereecke, E., D’Août, K., De Clercq, D., Van Elsacker, L., Aerts, P., 2003. Dynamic plantarpressure distribution during terrestrial locomotion of bonobos (Pan paniscus).Am. J. Phys. Anthropol. 120, 373e383.

Vereecke, E., D’Août, K., De Clercq, D., Van Elsacker, L., Aerts, P., 2004. The rela-tionship between speed, contact time and peak plantar pressure in terrestrialwalking of bonobos. Folia Primatol. 75, 266e278 (Basel).

Vereecke, E.E., D’Août, K., Payne, R., Aerts, P., 2005. Functional analysis of the footand ankle myology of gibbons and bonobos. J. Anat. 206, 453e476.

Vilensky, J.A., Gehlsen, G., 1984. Temporal gait parameters in humans and quad-rupeds: how do they change with speed? J. Hum. Move. Stud. 10, 175e188.

Weidenreich, F., 1923. Evolution of the human foot. Am. J. Phys. Anthropol. 6, 1e10.Wunderlich, R.E., 1999. Pedal Form and Plantar Pressure Distribution in

Anthropoid Primates. Ph.D. dissertation, State University of New York atStony Brook.

Zar, J.H., 1999. Biostatistical Analysis. Prentice Hall, Upper Saddle River, NJ.