Article pp.1-15 du Vol.27 n°1-2 (2015)

(1)

ARTICLE /ARTICLE

Comparative analysis of the coxo-femoral joint in hominids

in relation to their locomotor behaviour: study of joint correspondence and the degree of congruence between the three-dimensional axes

of the two joint components

Analyse comparative de l’articulation coxo-fémorale chez les hominidés en lien

avec leur répertoire locomoteur : étude de la congruence articulaire et du degré de concordance entre les axes tridimensionnels des deux composantes articulaires

N. Bonneau · C. Simonis · C. Tardieu

Received: 8 October 2013; Accepted: 5 May 2014

AbstractA comparative study of the hip joint in hominids has important implications for physical anthropology, as new insights could provide additional information for documenting changes associated with the acquisition of habitual bipedalism in the human lineage. In the literature, changes in the shape of both the proximal femur and the acetabular region in hominids have been investigated; however, few analyses have studied joint congruence and even fewer have attempted to quantify the degree of correspondence between the three-dimensional orientations of the two joint components. In humans, the hip joint has a lower degree of correspondence in a bipedal posture than would be expected for such a key joint. Because natural selection is dependent on inherited structure and trade-offs between several functions, it has been suggested in a previous study that the lower degree of correspondence might be partly due to the phylogenetic history of our species. In order to test this hypothesis, an extend sample of hominids was analysed. Our results show that joint architecture changes substantially according to the locomotor behaviour of the taxon considered. In orangutans, the considerable mobility of the lower limbs is ensured in part by a shallow acetabulum and a wide femoral

neck-shaft angle. In contrast, the hip joint of gorillas has to support heavy loads and thus needs stability: a deep acetabulum and a narrow femoral neck-shaft angle were observed.

Finally, our results provided evidence that the highest congruence between the three-dimensional orientations of the two joint components in all the studied genera, i.e.Homo, Pan,GorillaandPongo, was obtained in a quadrupedal posture, further suggesting that the shape of the human hip joint is constrained by its quadrupedal ancestry.

Keywords Hip joint · Bipedal gait and posture · Acetabulum · Femoral neck · Hominids

RésuméL’étude comparative de l’articulation de la hanche sur un échantillon d’hominidés trouve des applications importantes en anthropologie biologique puisqu’une meil- leure connaissance de cette articulation peut enrichir notre compréhension des changements morphologiques associés à l’acquisition de la bipédie permanente dans la lignée humaine. Dans la littérature, des changements morphologiques de l’extrémité proximale du fémur et de la région acét- abulaire ont été décrits, mais peu d’analyses se sont intéres- sées au paramètre de congruence articulaire et encore moins au niveau de concordance des axes tridimensionnels des deux composantes articulaires. Chez l’Homme, la hanche montre un plus faible niveau de concordance en posture bipède que ce qui aurait pu être attendu pour une articulation aussi centrale. L’action de la sélection naturelle étant néces- sairement tributaire de la nature de la structure héritée d’un ancêtre commun et des compromis qui s’établissent entre les différentes fonctions exercées par cette structure, nous avions suggéré dans un précédent travail que cette plus faible concordance soit imputée à l’histoire phylogénétique de

N. Bonneau (*)

UFR STAPS, Université Paris Sud XI, Bât. 335, 91405 Orsay cedex

e-mail : nbonneau@mnhn.fr N. Bonneau · C. Tardieu

UMR 7179, Muséum national d’Histoire naturelle/CNRS, CP 55 Pavillon d’Anatomie Comparée,

57, rue Cuvier 75231 Paris cedex 5 C. Simonis

Ministère de l’Education Nationale, DEPP, 75015 Paris DOI 10.1007/s13219-014-0118-x

(2)

notre espèce. Afin de tester cette hypothèse, une étude sur un échantillon d’hominidés a été menée. Nos résultats montrent que l’architecture articulaire diffère selon le répertoire locomoteur du taxon considéré. Chez les orangs-outans, la grande mobilité du membre inférieur est assurée en partie par une cavité acétabulaire peu profonde et un angle cervico-diaphysaire important. Au contraire, la hanche des gorilles doit supporter de fortes charges et nécessite plus de stabilité : une cavité acétabulaire profonde et un angle cervico-diaphysaire faible sont observés. Enfin, nos résultats mettent en évidence que le plus grand niveau de concordance des axes tridimensionnels des deux composantes articulaires de la hanche pour l’ensemble des genres consid- érés, c’est-à-direHomo, Pan,GorillaetPongo, est observée dans une posture quadrupède. Ce résultat appuie l’hypothèse d’une influence du passé quadrupède de nos ancêtres sur les orientations de l’articulation de la hanche actuelle.

Mots clésArticulation de la hanche · Locomotion et posture bipèdes · Acétabulum · Col fémoral · Hominidés

Introduction

Humans are characterized by a habitual and permanent bipedal gait and posture that places the centre of gravity of the whole body above the pelvis. During locomotion, the body mass is balanced first above one lower limb and then above the other, so that the one-legged stance represents a significant part of the human walking cycle [1]. Conse- quently, body weight is mainly supported by just one limb during locomotion in humans. In contrast, quadrupeds frequently support their weight on at least two legs, most often three, implying more even distribution of their body weight across the limbs [2-5]. The hip joint is a key element of the human locomotor system, as it plays an important role in supporting the body and transmitting forces between the erect trunk and lower limbs, especially during the one- legged stance. A comparative study of the hip joint among hominids therefore has important implications for physical anthropology, as new insights could provide additional information for documenting changes associated with the acquisition of habitual bipedalism in the human lineage.

Also of interest is the contribution of new material to the debate on the locomotor patterns of our most recent non- bipedal ancestors and additional knowledge to support a more accurate reconstruction of the locomotor behaviour of extinct species known only from fossils [6-15].

In nature, any joint has two main functions : it ensures transmission of the mechanical load, but also the ability or inability to perform particular movements [16]. The type of joint, and more specifically its exact structure, reflects the

need to ensure stability and mobility. A spheroid joint such as the hip joint provides three degrees of freedom of move- ment, resulting in high mobility while also ensuring stability.

However, in such joints the degree of interlocking of the ball in the socket is an important parameter that determines the predominance of stability over mobility or vice versa. Con- sequently, the degree of joint congruence is of prime importance in a functional approach to studies of the hip joint. In this study, a comparative analysis was made of the degree of congruence at the hip joint in a hominid sample comprising Homo, Pan,GorillaandPongo. The relationships between the degree of congruence and the preferred postures of each taxon during their habitual locomotor behaviour in their natural habitat were explored. The study of extant taxa is important to draw inferences on both the phylogenetic signal and the adaptive nature of the morphology of the hip in hominids. The different hominids studied here use a wide range of environments, terrestrial as well as arboreal, and have devel- oped a wide range of locomotor behaviour in response to these different environments [5,17]. In accordance with each type of behaviour, different mechanical demands are made on the hip joint, of which the anatomical configuration is expected to adapt through selection for efficient locomotion and the range of motion required. Despite the different postural and locomotor behaviour patterns displayed by these taxa, hominids are phylogenetically close, so that comparisons are less likely to be dependent on phylogenetic effects.

In addition to the degree of congruence, the orientations of the two neighbouring components of a joint, and more specifically the relationships between these two orientations during motion, is of prime importance to the way the joint functions. The orientations of the joint components reflect the direction of the forces acting upon it and thus have implications for load transmission. Joint correspondence is used to assess the relationships between the orientations of the neighbouring components of a joint in a given posture.

With full correspondence, the two orientations are aligned.

In humans, hip joint conrrespondence is lower in the bipedal posture than would be expected for such a key joint [18-20]:

both the acetabulum and the femoral neck present a normal anteversed orientation in a transverse plane, so that the anterior part of the femoral head is not fully covered by the acetabulum in the bipedal stance [18-21]. It has been suggested in the literature that the lack of full correspondence between the axes of the acetabulum and the femoral neck may result in premature wear on the joint and the appearance of osteoarthritis [18,22]. Because full correspondence of the axes of the two neighbouring components of the human hip joint is observed in a quadrupedal posture on an inclined substrate [20,23] we proposed in a previous study that the lower degree of correspondence might be partly due to the phylogenetic history of our species [19,20]. The hip joint of

(3)

modern humans evolved from an ancestral joint which was adapted to the specific locomotor demands of the last ancestor and the need to provide a wide enough passage during birth. Consequently, the evolution of the human hip joint is constrained by the properties of this pre-existing structure.

Moreover, the location of the hip joint in the pelvic girdle also implies that constraints related to the obstetrical role of the female pelvis have driven the evolution of its shape in response to an increase in head size and brain volume [24- 34]. Other functions known to influence pelvic shape are those related to the support of body viscera [25] and thermo- regulatory constraints [35]. To summarise, although natural selection has probably optimized the performance of the human hip joint for locomotor functions, its adaptation is also dependent, firstly, on the properties of the inherited structure and, secondly, on trade-offs with other functions.

Therefore, in order to test our previous hypothesis, the orientations of both the acetabulum and the femoral neck were analysed with special attention given to the degree of congruence between the two orientations in particular postures habitually used by the different taxa included in the study.

Differences in the shape of both the proximal femur and the acetabular region among hominids have been investigated in the literature [16,32,36-43]. However, few analyses have studied changes in the orientations of these two structures over the course of evolution [44] and even fewer have attempted to quantify their three-dimensional orientations.

The development of methods allowing quantification of the orientations of both the acetabulum and the femoral neck in three dimensions, and independent of anatomical reference planes [20,45], provides accurate tools for the present comparative analysis of the hip joint among hominids.

To summarise, this comparative analysis aims to quantify both hip joint correspondence and the degree of congruence between the three-dimensional orientations of both the acetabulum and the femoral neck during locomotion in the taxa Pan,GorillaandPongocompared withHomo sapiens. The joint parameters were studied in relation to the preferred pos-

tures of the different taxa during their habitual locomotor behaviour in their natural habitat.

Materials and Methods Materials

The study covered 66 specimens of modernHomo sapiens (Table 1) from the SIMON collection (housed at the Univer- sity of Geneva, Switzerland) and the collections of the National Museum of Natural History (Paris, France). The SIMON collection is composed of 20^th century skeletons collected from cemeteries in the Vaud canton [46]. The bones housed in the collections of the Paris Museum are 21^stcentury specimens from France.

An extended sample of non-human hominids was also analysed, including 48 specimens of Pan (troglodytes and paniscus), 36 of Gorilla (gorilla and beringei) and 22 of Pongo pygmaeus(Table 1). The bones of these specimens were housed in the National Museum of Natural History (Paris, France), the Anthropological Institute and Museum (University of Zurich, Switzerland), the Belgian Royal Insti- tute of Natural Sciences (Brussels, Belgium) and the Royal Museum of Central Africa (Tervuren, Belgium).

Data acquisition

The bones were immobilised in a clamp and their three- dimensional coordinates (x, y, z) were recorded in a milli- metric orthonormal reference system using Microscribe^®G2 (Immersion, France), which is accurate to ± 0.38 mm according to the manufacturer.

The three-dimensional coordinates (x, y, z) of six homologous landmarks on the hip bones (Table 2, Fig. 1) were digitized. In addition, acquisitions were also performed on the acetabular rim using Microscribe programmed to take coordinates one millimetre apart. First, successive points

Table 1 Summary of the material used in the comparative study of the hip joint in hominids /Bilan du matériel utilisé pour l’étude comparative de l’articulation de la hanche chez les hominidés.

Taxon N°specimens Femora Hip bones Museum*

Total Right Left Total Right Left

Homo sapiens 66 91 43 48 99 52 47 UG, MNHN

Pan spp. 48 72 36 36 91 45 46 MNHN, AIM, IRSNB, RMCA

Gorilla spp. 36 48 23 25 60 30 30 MNHN, AIM, IRSNB, RMCA

Pongo pygmaeus 22 28 15 13 37 18 19 MNHN, AIM, IRSNB

* UG = University of Geneva (Switzerland), MNHN = Muséum national d’Histoire naturelle et Musée de l’Homme (Paris, France), AIM = Anthropological Institute and Museum (Zürich, Switzerland), IRSNB = Institut Royal des Sciences Naturelles de Belgique (Brussels, Belgium), RMCA = Royal Museum for Central Africa (Tervuren, Belgium).

(4)

were recorded along the entire acetabular rim, i.e. from the anterior horn tip to the posterior horn tip [47]. Next, the anterior and posterior parts of the acetabular rim were digitized according to the anatomical description proposed by Bonneau and collaborators [47]. These two parts were delimited by the two notches identified in the rim: one on its upper part, at the intersection of the anterior edge of the ilium with the acetabular rim, and the second on its posterior part, where the main axis of elongation of the ischium meets the osseous acetabular rim (Fig. 1 in [19]). Finally, successive points were acquired along the edges of the facies lunata, the characteristically horseshoe-shaped articular surface of the acetabulum.

For the femora, the three-dimensional coordinates of eight homologous landmarks (Table 2, Fig. 1) were digitized. In addition, using the stylus of the Microscribe, which was programmed to take three-dimensional coordinates one millimetre apart, the outer surface of the femoral neck between the edge of the articular surface of the femoral head and the intertrochanteric lines was traced, with 800 to 1600 points recorded depending on the genus. Using the

same method, three-dimensional coordinates were recorded on the surface of the femoral shaft between the two metaphyses.

The landmarks acquired on both hip bones and femora were chosen for their minimal observer-induced measure- ment error [20,45,48] and their maxima and homogeneous dispersion relative to the overall bone volume.

Level of congruence at the hip joint

All analyses were performed using the R graphics and statis- tics package v.2.11.0 [49]. In order to assess the degree of congruence at the hip joint in hominids, the percentage of a sphere enclosed by the acetabular socket was computed for each genus. The three-dimensional coordinates acquired along the edges of thefacies lunatawere used to compute the geometrical centre and radius of the articular surface of each acetabulum based on a spherical regression. The three- dimensional coordinates of the acetabular centre were projected onto a regression plane computed from the points acquired along the entire acetabular rim. The distance between the acetabular centre and its projection was recorded (d in Fig. 2) and this value was subtracted or added to the value of the radius of the acetabulum, depending on whether the acetabulum forms less or more than a hemisphere. This estimation of acetabular depth was used to calculate the percentage of a sphere formed by the acetabular socket. The depth was multiplied by 100 and divided by the radius of the acetabulum. In parallel, the angle formed between the anterior and the posterior parts of the acetabular rim was calculated, because the morphology of the osseous acetabular rim, which is known to be irregular [47,50-54], influences the nature of joint congruence since a wider angle between the anterior and the posterior parts of the rim could result in less coverage of the femoral head. A regression plane, using the least squares method, was computed from the points acquired on the anterior part of the osseous acetabular rim on the one hand, and on the posterior part of the rim on the other hand. The angle formed by the two direction vectors of the two regression planes corresponding to the orientation of the anterior and posterior parts of the acetabular rim was computed.

The femoral neck-shaft angle

The femoral neck-shaft angle was compared across the entire sample of hominid femora. For each femur, the three- dimensional coordinates characterizing the femoral neck surface were used to determine the three-dimensional orientation of the femoral neck, based on the method described by Bonneau and collaborators [45]. This innovative method uses successive cross-sectional ellipses to model the femoral neck surface, allowing accurate determination of its three- Table 2 List of homologous landmarks used in the study /Liste

des points homologues utilisés dans l’étude.

No Definition HIP BONE

1 Upper point of the pubic symphysis 2 Upper anterior iliac spine

3 Spinea limitans: tubercle formed by the lower posterior border of the insertion of thequadratum lumborum muscle, at the lateral border of the iliac crest 4 Lower posterior iliac spine

5 Uppermost posterior point of the ischial tuberosity, i.e.

upper border of the insertion of thesemimembranosus muscle

6 Scalenion: tangent point of the line between the acetabular centre and the ventral border of the sacroiliac joint [77]

FEMUR

1 Pretrochanteric tubercle, at the iliofemoral ligament attachment

2 Lower anterior point of insertion of thegluteus minimus 3 Uppermost point of thepiriformismuscle attachment 4 Upper posterior point of the posterior intertrochanteric

crest

5 Posterior base of the medial condyle 6 Posterior base of the lateral condyle

7 Maximum curvature of the central part of the trochlea (frontal view)

8 Maximum curvature of the distal articular surface between the two condyles

(5)

dimensional orientation based on the successive centres of the different ellipses. The use of successive ellipses gives great flexibility to the model, which can match circular or elliptical cylinders as well as circular or elliptical cones, thus justifying its application for all the different hominoids analysed. In addition, the three-dimensional coordinates

characterizing the femoral shaft surface were used to determine the three-dimensional orientation of the main axis of the shaft based on a circular cylindrical regression. The angle between the femoral shaft axis and the femoral neck axis was computed to obtain the femoral neck-shaft angle (FNA in Fig. 2).

Fig. 1 Homologous landmarks used in the study and further described in Table 2. The scans of the bones were performed using a Stereo- scan Breukmann® surface scanner. A–lateral view of the right hip bone; B –medial view of the right hip bone; C–frontal view of the proximal part of the right femur; D–posterior view of the distal part of the right femur; E–proximal view of the right femur;

F–distal view of the right femur /Points homologues utilisés dans l’étude et détaillés dans leTableau 2.Les scanners des os ont été réalisés grâce à un Scanner surfacique Breaukmann®. A–vue latérale de l’os coxal droit ; B –vue médiale de l’os coxal droit ; C–vue frontale de l’extrémité proximale du fémur droit ; D–vue postérieure de l’extrémité distale du fémur droit ; E–vue proximale du fémur droit ; F–vue distale du fémur droit.

(6)

Fig. 2 The comparative analysis of the hip joint was based on two parameters: the femoral neck-shaft angle and the percentage of a sphere formed by the acetabular socket. The femoral neck-shaft angle (FNA) corresponds to the angle between the three- dimensional axis of the femoral neck and the three-dimensional axis of the femoral shaft. Relative to the acetabulum, the three- dimensional coordinates of the acetabular centre (O) were projected onto the regression plane (P) computed from the points acquired on the total acetabular rim. The distance between the acetabular centre and its projection was recorded (d) and this value was subtracted or added to the value of the radius of the acetabulum, depending on whether the acetabulum forms less or more than a hemisphere. In the picture, the acetabulum forms less than a hemisphere, so that the value of d was subtracted from the value of R. This estimation of acetabular depth was used to calculate the percentage of a sphere formed by the acetabular socket. The depth was multiplied by 100 and divided by the radius of the acetabulum /Deux paramètres ont été utilisés dans cette étude comparative de la hanche : l’angle cervico-diaphysaire du fémur et le pourcentage de sphère formé par la cavité acétabulaire. L’angle cervico-diaphysaire du fémur (FNA) correspond à l’angle formé entre l’axe tridimensionnel du col fémoral et l’axe tridimensionnel de la diaphyse fémorale. Concer- nant l’acétabulum, les coordonnées tridimensionnelles du centre acétabulaire (O) ont été projetées sur le plan de régression (P) calculé à partir des points numérisés sur le pourtour total du sourcil acétabulaire. La distance entre le centre acétabulaire et la projection de ce centre sur P a été calculé (d) et cette valeur a été soustraite ou additionnée à la valeur du rayon de la sphère acétabulaire, selon que l’acétabulum faisait plus ou moins une demi-sphère. Dans cette illustration, l’acetabulum représenté forme moins d’une demi-sphère, donc la valeur de d a été soustraite à celle de R. Cette estimation de la profondeur acétabulaire a été ensuite utilisée pour calculer le pourcentage de sphère formé par la cavité acétabulaire. La profondeur à été multipliée par 100 et divisée par le rayon de l’acétabulum.

(7)

Degree of correspondence between the axes

of the two hip joint components in different postures

The differences in the three-dimensional orientations of the acetabulum and the femoral neck in the hominid specimens were explored in the context of the habitual postures adopted by each taxon during their locomotor behaviour in their natural habitat. The three-dimensional orientations of both the femur and the hip bone in these habitual postures were used to establish the reference area for comparison between the different hominids, rather than a mathematical superimposition of bones based on homologous landmarks. While superimposition of the bones of various taxa is used to analyse inter-specific differences in shape, there is no point in using a functional approach, for two reasons. First, given that the relative position and orientation of both the femora and the hip bones depends on the specific locomotor behaviour of the hominid considered, the functional frontal and transverse planes of the two bones may not correspond across taxa. Moreover, these planes do not necessarily correspond to the anatomical planes of reference used to describe these bones in comparative anatomy. For example, Lewin- neck’s plane calculated from the two upper anterior iliac spines and the upper point of the pubic symphysis is known to be vertical in the bipedal posture of humans. How- ever, this plane makes no sense in chimpanzees as the pelvic girdle is more horizontal, being associated with their quadrupedal walking. Thus, the vertical forces acting on the acetabulum in these two taxa do not have the same three- dimensional orientation in the pelvic cavity. Secondly, as superimposition is based on homologous landmarks, it is necessarily dependent on the total conformation of the bones. However, profoundly different orientations of the three parts of the hip bone are known to exist between hominids. A different orientation of the pubic ramus, for example, influences superimposition as the variation in the location of the upper point of the pubic symphysis will be distributed between the eight homologous landmarks of the hip bone because of the mean squared approach. The orientation of the acetabulum is thus strongly influenced by the total shape of the bone, as important differences between taxa are identified. For these reasons, we decided to analyse the three-dimensional orientation of the acetabulum and the femoral neck, and their relationships, in postures that are specific to each taxon.

Based on data available in the literature [5,17,55-61], the most habitual postures adopted by the different taxa during locomotion in their natural habitat were determined. The bipedal posture was used to illustrate the locomotor behaviour of Homo sapiens. As maximum congruence of the human hip joint has been described in a quadrupedal posture on an inclined substrate [20,23], this posture was also considered. As gorillas (Gorilla) are among the most terrestrial

of all primates, they were studied in their quadrupedal knuckle-walking posture [5]. Chimpanzees and bonobos (Pan) are known to use more arboreal locomotion than gorillas. Consequently, both the quadrupedal knuckle-walking posture used to travel on the ground and the suspensory posture used to move among trees were used to illustrate locomotor behaviour inPan. Orangutans (Pongo) are known for their suspensory behaviour. Two postures were chosen to represent the slow quadrumanous climbing that they habitually use during arboreal locomotion. As adult males frequently travel to the ground [5], the terrestrial quadrupedal posture was also studied.

For each taxon, two separate partial superimpositions of the hip bones and femora were performed using a custom- designed function of the Rmorph library for R [20,45,48,62].

The left-side bones were made symmetrical using left-right matching symmetry to obtain right and left bones in a com- parable space. The superimposition process, based on a Gen- eralized Procrustes Analysis [63,64], involved a scaling step followed by translations and rigid three-dimensional rotations of the bones using the homologous landmarks. The three-dimensional coordinates acquired on the osseous acetabular rims and femoral neck surfaces passively followed the translations and rotations calculated from, respectively, the six homologous landmarks acquired on the hip bones and the eight homologous landmarks acquired on the femora. To summarise, this alignment removes the effect of variation in position and orientation from the three-dimensional coordinates acquired on both the acetabular rims and the femoral neck surfaces.

Mean shapes for each taxon were computed separately from the homologous landmarks acquired on the hip bones and the femora. The mean shapes obtained of the femora and the hip bones were reoriented to the standard positions they have in the habitual posture(s) chosen to describe the locomotor behaviour of the taxa considered. More specifically, the coordinates of the mean shapes of both hip bones and femora for each taxon were transformed mathematically in order to rearticulate the two components of the hip as they appear in the posture analysed. This mathematical transfor- mation of the mean shapes was performed once only for each taxon and each position to avoid the intra- and inter- variability that would be introduced if the reorientation is performed for each individual before the superimposition process. Using this protocol, a general posture representing a locomotor behaviour pattern was thus analysed. After reorientation, the three-dimensional coordinates characterizing the femoral neck surfaces were used to determine the three-dimensional orientation of each femoral neck [45].

The successive points characterizing the acetabular rims were used to determine the three-dimensional orientation of each acetabulum [47]. Mean vectors were computed for the three-dimensional orientations of both the acetabulum

(8)

and the femoral neck. The ranges of variation of the two orientations are illustrated. Joint correspondence is used to estimate the relationships between the orientations of the neighbouring components of the joint in each posture, i.e.

in a given posture, the orientation of one component of the joint was observed in relation to the orientation of the other component. Full correspondence occurs when the two orientations are in alignment.

Results

Our results are summarized in Table 3 and figures 3 to 5.

Joint congruence was assessed based on two parameters, the percentage of a sphere formed by the acetabular cavity and the angle between the anterior and the posterior parts of the osseous acetabular rim. Our results provide evidence that the percentage of a sphere formed by the acetabulum in orangutans is smaller than in the other genera studied (33.77% on average, Table 3), with a wide angle between the anterior and posterior parts of the acetabular rim (41.75° on average).

In contrast, a percentage of a sphere equal to 42.84% on average was obtained for gorillas, also with a wide angle between the anterior and posterior parts of the rim (42.62°

on average). In chimpanzees and bonobos, the mean percentage of a sphere is 40.15% and the mean angle between the anterior and posterior parts of the rim is 40.93°. Finally, in humans, the percentage of a sphere is close to the value obtained for chimpanzees and bonobos (40.99% on average) while the angle formed by the anterior and the posterior parts of the osseous rim was narrower than in other hominids (21.75° on average).

Concerning the femoral neck shaft angle, from which the orientation of the femoral neck is estimated, the maximum value was calculated in orangutans (136.5° on average) while the narrowest angle was calculated for gorillas (120.0°

on average). This angle is equal on average to 128.9° in humans and 126.5° in chimpanzees and bonobos.

Joint correspondence is used to estimate the relationships between the orientations of the neighbouring components of the joint in a given posture i.e., a posture in which the orientation of one component of the joint is observed in relation to the orientation of the other component. Full correspondence occurs when the two orientations are aligned, i.e. when the angle formed between the two axes is equal to 180°. In the habitual bipedal posture, the correspondence between the two axes of the acetabulum and the femoral neck in the artic- ulated hip process was lower than expected for a strong joint such as the hip joint. Regarding the mean orientations of both the acetabulum and the femoral neck in the bipedal posture, correspondence between the two axes was high in the frontal view (170°; Fig. 3) but quite low in the transverse plane (144°; Fig. 3), where the orientation of both the acetabulum and the femoral neck was anteversed, resulting in lower correspondence between the orientations of the two neighbouring components. Our results show that the highest correspondence between the two hip joint components for all the genera studied, i.e.Homo,Pan,GorillaandPongo, was obtained in quadrupedal postures (Fig. 4). Since this was predicted by Kapandji [23] and confirmed in this study, the maximum correspondence between the axes of the two joint components in humans was observed in a quadrupedal posture on an inclined substrate, when the angle formed by the two axes in the transverse plane reached 176° (Fig. 4C). In gorillas, 174° and 159° of correspondence between the axes of the two joint components were measured in the frontal and transverse planes respectively, implying a high degree of correspondence in the quadrupedal knuckle-walking posture (Fig. 4A). In chimpanzees and bonobos (Pan), angles equal to 169° and 170° in the frontal and transverse planes were measured in the quadrupedal knuckle-walking posture (Fig. 4B), while these same angles were equal to 175° and

Table 3 Results of the comparative analysis of femoral and acetabular parameters between the different hominid genera /Résultats de l’étude comparative des paramètres fémoraux et acétabulaires obtenus sur un échantillon d’hominidés.

ACETABULUM

Homo sapiens Pan spp. Gorilla spp. Pongo pygm.

Percentage of sphere formed by the acetabular socket (%)

40.99 ± 3.46 (33.24-53.21)

40.15 ± 4.08 (31.96-52.28)

42.84 ± 5.51 (34.47-52.20)

33.77 ± 5.75 (23.09-42.53) Angle between the anterior and posterior

parts of the acetabular rim

21.75 ± 8.92 (10.15-41.00)

40.93 ± 7.41 (31.84-56.38)

42.62 ± 8.02 (33.15-58.76)

41.75 ± 7.66 (33.09-50.31) FEMUR

Homo sapiens Pan spp. Gorilla spp. Pongo pygm.

Femoral neck-shaft angle (°) 128.9 ± 5.90 (118.5-140.7)

126.5 ± 6.45 (116.8-138.3)

120.0 ± 6.07 (109.8-130.1)

136.5 ± 9.03 (121.2-151.3)

(9)

Fig. 3 The coordinates of the mean configurations of both the hip bones and the femora were transformed mathematically in order to rearticulate the two components of the hip as they appear in a standard bipedal posture. The mean axes of the acetabulum and the femoral neck (red lines), computed from the human sample, were illustrated in a frontal view and a transverse view. The dotted lines show the range of variation of the three-dimensional orientations of the acetabulum and the femoral neck. In a bipedal posture, the upper anterior part of the femoral head is uncovered. Mainly in the transverse plane, both the acetabulum and the femoral neck have an anteversed orientation resulting in lower congruence than would be expected for a key joint such as the hip joint /Les coordonnées des conforma- tions moyennes des os coxaux et des fémurs ont été mathématiquement réorientées afin de réarticuler les deux composantes de la hanche telles qu’elles sont orientées dans une posture bipède standard. Les axes moyens de l’acétabulum et du col fémoral (lignes continues rouges), calculés à partir de l’échantillon humain, sont représentés dans une vue frontale et une vue transverse. Des lignes pointillées illustrent la variation inter-individuelle des orientations tridimensionnelles de l’acétabulum et du col fémoral dans notre échantillon.

Dans une posture bipède, la partie antéro-supérieure de la tête fémorale est découverte. Dans le plan transverse, l’acétabulum et le col fémoral ont des orientations antéversées responsables de la plus faible concordance des axes que ce qui pourrait être attendu pour une articulation centrale telle que la hanche.

(10)

Fig. 4 Quadrupedal postures of the four genera studied, i.e. Gorilla(A),Pan(B),Homo(C) andPongo(D). The drawings illustrate the general posture. Personal scans made by Laurent Puymerail or scans provided by the Digital Morphology Museum (Kupri) are used to illustrate the orientations of the right hip bone and femur. The mean axes of the acetabulum and the femoral neck (full lines) are illustrated in both frontal and transverse views using the respective samples of the taxa analysed. The dotted lines illustrate the range of variation of the three-dimensional orientations of the acetabulum and the femoral neck for each taxon. A– Gorillas (Gorilla) are among the most terrestrial of all primates and are illustrated in their quadrupedal knuckle-walking posture. B–Chimpanzees and bonobos (Pan) are known to travel on the ground using a quadrupedal knuckle-walking posture. C – As the maximum correspondence between the two components of the human hip joint was described in a quadrupedal posture on an inclined substrate [20,23], this posture was analysed. D–Although orangutans (Pongo) are known for their suspensory behaviour, adult males frequently travel on the ground. The terrestrial quadrupedal posture is illustrated here /Des postures quadrupèdes utilisées par les quatre genres étudiés, c’est-à-direGorilla(A), Pan(B), Homo(C) etPongo(D), ont été représentées. Le croquis illustre la posture générale de l’animal. Soit des scanners personnels de Laurent Puymerail soit des scanners provenant du Digital Morphology Museum (Kupri) ont été utilisés pour illustrer les orientations de l’os coxal et du fémur droit. Les axes moyens de l’acétabulum et du col fémoral sont représentés (lignes continues) dans le plan frontal et le plan transverse à partir des données acquises respectivement sur chaque taxon. Des lignes pointillées illustrent la variation inter-individuelle des orientations tridimensionnelles de l’acétabulum et du col fémoral de chaque taxon. A–Les gorilles (Gorilla) repré- sentent un des genres les plus terrestres parmi les primates. Ils ont donc été représentés dans leur posture quadrupède en knuckle- walking. B– Les chimpanzés et les bonobos (Pan) voyagent au sol en utilisant la posture quadrupède knuckle-walking. C – Dans la mesure où le maximum de concordance des axes des deux composantes de la hanche humaine a été décrit en posture quadru- pède sur un plan incliné [20,23], cette posture a été analysée. D–Bien que les orangs-outangs (Pongo) passent une grande partie de leur temps dans les arbres, les mâles adultes voyagent fréquemment au sol. A ce titre la posture quadrupède au sol a également été représentée.

(11)

152° in the arboreal posture observed in this study (Fig. 5A).

In orangutans, correspondence between the axes of the two hip joint components reached 163° and 161° in the frontal and transverse planes of the quadrupedal posture (Fig. 4D), while the same angles were equal to 176° and 137° during slow quadrumanous climbing (Fig. 5B), and 161° and 134°

in the posture requiring a high degree of abduction of the lower limbs (Fig. 5C).

Discussion

Our results provide new quantitative data on hip joint congruence and on the degree of correspondence between the three-dimensional orientations of both the acetabulum and the femoral neck during locomotion in hominids. The morphology of the hip joint is a compromise between the need for mobility and for stability, and the former or the latter are

Fig. 5 Arboreal postures used by chimpanzees and bonobos (Pan) and orangutans (Pongo) in their natural habitat. The drawings illustrate the general posture. Personal scans made by Laurent Puymerail or scans provided by the Digital Morphology Museum (Kupri) are used to illustrate the orientation of the right hip bone and femur. The mean axes of the acetabulum and the femoral neck (full lines) are illustrated in both frontal and transverse views using the respective samples of the taxa analysed. The dotted lines illustrate the range of variation of the three-dimensional orientations of the acetabulum and the femoral neck for each taxon. A–Chimpanzees and bonobos (Pan) are known to use arboreal locomotion more frequently than gorillas (Gorilla). The suspensory posture used to move around in the trees is therefore illustrated here. B–Orangutans (Pongo) are known for their suspensory behaviour. This posture was therefore chosen to represent the slow quadrumanous climbing that they habitually use during arboreal locomotion. C–Orangutans are known for their extreme specialization for arboreal suspensory locomotion, associated with very high mobility of the hind limbs, and in particular for their ability to abduct the femur to a considerable degree /Les postures arboricoles utilisées par les chimpanzés et les bonobos (Pan) et les orangs-outangs (Pongo) dans leur habitat naturel sont représentées. Le croquis illustre la posture générale de l’animal. Soit des scanners personnels de Laurent Puymerail soit des scanners provenant du Digital Morphology Museum (Kupri) ont été utilisés pour illustrer les orientations de l’os coxal et du fémur droit. Les axes moyens de l’acétabulum et du col fémoral sont représentés (lignes continues) dans le plan frontal et le plan transverse à partir des données acquises respectivement sur chaque taxon. Des lignes pointil- lées illustrent la variation inter-individuelle des orientations tridimensionnelles de l’acétabulum et du col fémoral pour les différents taxa. A–Les chimpanzés et bonobos (Pan) utilisent plus fréquemment la locomotion arboricole que les gorilles. De ce fait, une position en posture suspendue a été illustrée. B– Les orangs-outangs (Pongo) utilisent très fréquemment la brachiation. La posture choisie dans cette illustration représente une position arboricole dans laquelle l’animal utilise les quatre membres. C–Les orangs-outangs sont connus pour leur extrême spécialisation en locomotion arboricole, associée à une grande mobilité des membres inférieurs ; en particulier leur capacité à l’abduction dans de forts niveaux d’amplitudes.

(12)

favoured according to the locomotor behaviour of the animal. Orangutans are known for their extreme specialization for arboreal suspensory locomotion, associated with very high mobility of the hindlimbs [5,58-61]. Several features demonstrate the very high mobility of the orangutan hip joint. It is known that orangutans, given their femoral pro- portions, have small diaphyseal and femoral neck dimensions compared to other hominids, a characterististic associated with reduced loading of the lower limbs due to their arboreal suspensory locomotion [39,65]. However, these smaller dimensions are not reflected in the size of the femoral head, since the isometric relationship described between femoral head size and body mass in non-human hominids remains [16,39]. This results in a larger femoral head size relatively to the femoral neck size, a configuration that favours a greater range of motion [39]. Moreover, theliga- mentum teresis absent and the articular surface of the femoral head forms a larger proportion of a sphere than in the other taxa (77% for orangutans vs. 71% for chimpanzees and gorillas [39-40]), both increasing the joint excursion.

Our results concur with the data available in the literature.

First, the high value of the mean femoral neck-shaft angle obtained in our sample of orangutans is in accordance with their ability to abduct the femur to a considerable degree, as illustrated in Figure 5C. Without this femoral neck orientation, such a posture would not be possible, as the femoral neck axis moves beyond the acetabular rim and the femoral neck surface would thus rapidly abut against the acetabular rim. Joint mobility is also facilitated by the small percentage of a sphere formed by the acetabular socket (34% on average), i.e. the low degree of joint congruence, which is in accordance with the small value of acetabular depth in orangutans already observed in the literature [39,66]. Despite the small amount of coverage provided by the acetabular socket for the femoral head, orangutans have no problem with hip stability as they load their hindlimbs less than the other taxa, as confirmed by the homogeneous organisation of the tra- beculae observed in their lumbar vertebral bodies [67].

In contrast, in gorillas, which are among the most terrestrial of all primates [5,55], the hip joint is very much involved in supporting their body mass. Moreover, adult males especially may reach a large body mass, thus increasing the load on the hip joint. Consequently, a stable hip joint is required. This need for stability may explain the high percentage of a sphere formed by the acetabular socket (43% on average) allowing substantial coverage of the femoral head by the acetabular socket. This deeper acetabulum, also noted in the literature [39-40,66], is accompanied by a large femoral head in accordance with the large body mass of the species. The isometric relationship described between the femoral head size and the body mass in non-human hominids reflects the increase in the weight-bearing surface, resulting in a smaller unit load [39,68] in species with a large body mass. The high corre-

spondence of the three-dimensional orientations of both the acetabulum and the femoral neck in the quadrupedal knuckle- walking posture in gorillas is also an indication of hip joint stability in this posture (Fig. 4A). The narrow mean femoral neck-shaft angle in our sample of gorillas suggests limited hip abduction, which is again in accordance with their terrestrial life-style.

The results obtained for the genusPansuggest an intermediate morphology between the stable hip joint of gorillas and the very mobile joint of orangutans. This intermediate morphology is in accordance with the locomotor behaviour of species in this genus. Both chimpanzees and bonobos are known to divide their time between the ground (Fig. 4B) and trees (Fig. 5A), which implies a need for joint mobility during arboreal locomotion and joint stability during terrestrial locomotion. However, their hip joint morphology does not allow as much mobility as observed for the orangutan, and less hip stability is required in comparison to gorillas due to their smaller body mass.

Hip joint morphology in modern humans is specific in several respects pertaining to their bipedal gait and posture.

First, while an isometric relationship has been described between femoral head size and body mass in non-human hominids [16,39], humans have a larger femoral head size for their body mass [39,43,69]. This is probably due to the fact that during the one-legged stance the body mass is supported by only one limb [70]. In contrast with orangutans, this large femoral head is associated with a deep acetabulum, since the human acetabulum has a mean percentage of a sphere equal to 41%. Rather than an adaptation for joint excursion as proposed for orangutans, the human configuration is more likely to reflect an adaptation for joint stability due to the increased load on the hip joint resulting from bipedal locomotion. Concerning the specific orientation of the femoral head, this can be explained by the considerable demands made on the abductor apparatus to prevent pelvic drop during bipedal locomotion [71-73]. Thus, while the terrestrial behaviour of humans, associated with high load bearing, and the reduced need for abduction, could lead us to predict a small femoral neck-shaft angle, the angle is equal to 129° on average, which is wider than that observed in both chimpanzees and bonobos. This orientation provides the abductors with the greater leverage required to stabilize the hip during walking and to increase the effectiveness of lateral balance control. Finally, a specific human feature observed in our data is the smaller angle between the planes formed by the anterior and the posterior parts of the acetabulum. We had previously provided evidence that the acetabular rim is better represented by two planes than by a single plane [47]. This irregularity in the orientation of the different parts of the osseous acetabular rim was interpreted on the basis of our observations of hip development in humans, where each flange of the triradiate cartilage is subject to

(13)

different forces, resulting in a different orientation for each part of the acetabular rim [47]. We therefore suggested that the orientation of the total acetabular rim depends on the intrinsic morphology of the pelvic girdle, which is in accordance with the forces transmitted from the upper body to the lower limbs through the pelvic girdle. The wider angle formed by the anterior and the posterior part of the acetabular rim calculated for the three other great apes,Gorilla,Pan andPongo, may be related to the different morphology of the pelvis in non-human great apes and humans, since adaptation of the pelvis to a bipedal gait and posture has resulted in a profound change in the relative orientation of the ilium, the pubis and the ischium. We suggest here that these changes may explain the different relative orientations of the different parts of the osseous acetabular rim between the taxa studied. Because this angle is a good estimator of the overall morphology of the pelvis and the acetabular region is generally well preserved, this feature can be a use- ful parameter to help reconstruct the locomotor behaviour of fossil species. However, as noted by Ruff and Runestad [16], the hip joint has to be considered as a complex of characters, and an interpretation based on one feature only could lead to errors.

This study of the relationships between the three-dimensional orientation of the acetabulum and the three-dimensional orientation of the femoral neck provides evidence that the highest correspondence between the three-dimensional axes of the two components of the hip joint in the four hominid taxa occurs in the quadrupedal posture (Fig. 4A-D).

Although this result was predicted for the genera Gorilla andPan, which spend a significant proportion of their time in this posture, this result was somewhat surprising for orangutans, which are more arboreal. Most surprising is the result for humans, which never practice quadrupedal locomotion as adults yet show the highest correspondence between the axes of the two components in the quadrupedal posture (Fig. 4C).

These observations suggest that the shape of the human hip joint partly reflects the phylogenetic history of the human species, further supporting our initial hypothesis. The fact that some morphological traits may have been retained from common ancestors in descendant taxa is a consequence of the nature of biological evolution [74]. Although many evolutionary traits in human hip joint morphology are known to be functionally involved in habitual bipedal locomotion [16,32,36-43], our results provide evidence that the quadrupedal origin of our species might be recorded in the three- dimensional orientations of the two neighbouring components of the human hip joint. This record of ancestral locomotor patterns in the three-dimensional orientations of the two components of the human hip joint is understood regarding the evolutionary processes. As suggested in the introduction, although natural selection optimizes the performance of the human hip joint according to locomotion, its adaptation is

also dependent on, first, the physical constraints of the inherited structure and, secondly, trade-offs with other functions.

As the quadrupedal posture is the one in which the highest correspondence between the three-dimensional orientations of the two components of the hip joint was observed in the four taxa studied, even though none of these taxa is exclu- sively quadrupedal, it can be assumed that it has been retained from a quadrupedal common ancestor that diverged prior to the orangutan/African ape clade, i.e. before the divergence of Hominidae. Although our data point to a quadrupedal common ancestor, it is difficult to distinguish between the different kinds of quadrupedal locomotion and, therefore to enlighten the ongoing discussion on whether the locomotor system of our last common ancestor was adapted for either orthograde arboreal locomotion or pronograde terrestrial quadrupedalism [6-15]. It would be interesting to analyse additional postures, including postures adopted by the Hylo- batidae. The locomotion ofHylobatesis known to be versatile [74], as they use brachiating as well as leaping and bipedal behaviour [76]. An extended study would provide more information on the entire Hominoïdea superfamily, and would be of particular value for comparisons with other primates including humans.

AcknowledgmentsThanks to the different people in charge of the collections we visited for the study: Geneviève Per- réard (curator of the SIMON collection at the University of Geneva), Philippe Mennecier (Musée de l’Homme, Paris), Christine Lefèvre (National Natural History Museum, Paris), Marcia Ponce de León and Christoph Zollikofer (Anthropological Institute and Museum, University of Zür- ich), Georges Lenglet (Belgian Royal Institute of Natural Sciences), and Emmanuel Gilissen (Tervuren Museum on Central Africa). Thanks also to Anthony Herrel for his comments and corrections of this paper. Finally, special thanks to Laurent Puymerail for the scans of hominid hip bones and femora used for the illustrations. Financial support from the Société d’Anatomie de Parisand the Action Transversale Muséum ‘Formes possibles, formes réalisées’ programme run by the National Natural History Museum (Paris) are gratefully acknowledged.

Conflict of interest :The authors do not have any conflict of interest to declare

References

1. Inman VT, Ralston HJ, Todd F (1981) Human walking. Williams

& Wilkins, Baltimore

2. Gray J (1944) Studies of the mechanics of tetrapod skeleton.

J Exp Biol 20:88–116

3. Reynolds TR (1985) Mechanics of increase support of weight by the hindlimbs in primates. Am J Phys Anthropol 67:335–49

(14)

4. Reynolds TR (1985) Stresses on the limbs of quadrupedal primates. Am J Phys Anthropol 67:351–362

5. Fleagle JG (1999) Primate Adaptation and Evolution. Academic Press, London

6. Prost JH (1980) Origin of bipedalism. Am J Phys Anthropol 52:175–89

7. Begun DR (1993) Knuckle-walking ancestor. Science 259:294 8. Gebo DL (1996) Climbing, brachiation, and terrestrial quadru-

pedalism: historical precursors of hominid bipedalism. Am J Phys Anthropol 101:55–92

9. Fleagle JG, Stern JT, Jungers WL, et al (1981) Climbing: a biomechanical link with brachiation and bipedalism. Symp Zool Soc Lond 48:359–75

10. Richmond BG, Begun DR, Strait DV (2001) Origin of human bipedalism: knucckle-walking hypothesis revisited. Yearbook Phys Anthrop Suppl 33:70–105

11. Thorpe SKS, Holder RL, Crompton RH (2007) Origin of human bipedalism as an adaptation for locomotion on flexible branches.

Science 316:1328–31

12. Crompton RH, Vereecke EE, Thorpe KS (2008) Locomotion and posture from the common hominoid ancestor to fully modern hominins, with special reference to the last common panin/homi- nin ancestor. J Anat 212:501–43

13. White TD, Asfaw B, Beyene Y, et al (2009)Ardipithecus rami- dusand the Paleobiology of early hominids. Science 326:64–86 14. Sarmiento EE (2010) Comments on the Paleobiology and classi-

fication ofArdipithecus ramidus. Science 328:1105.

15. Crompton RH, Sellers WI, Thorpe SKS (2010) Arboreality, terrestriality and bipedalism. Phil Trans R Soc B 365:3301–14 16. Ruff C, Runestad JA (1992) Primate limb bone structural adapta-

tions. Annu Rev Anthropol 21:407–33

17. Hunt KD, Cant JGH, Gebo DL, et al (1996) Standardized description of primate locomotor and postural modes, Primates 37:363–87

18. Reikerås O, Bjerkreim I, Kolbenstvedt A (1983) Anteversion of the acetabulum and femoral neck in normals and in patients with osteoarthritis of the hip. Acta Orthop Scand 54:18–23

19. Bonneau N (2013) Three-dimensional analysis of the pelvi-femoral complex: its mechanism and development. Evolutionary and clini- cal implications. Bull Mém Soc Anthropol Paris 25:208–16 20. Bonneau N, Baylac M, Gagey O, Tardieu C (2014) Functional

integrative analysis of the human hip joint: the three- dimensional orientation of the acetabulum and its relation with the orientation of the femoral neck. J Hum Evol 69:55–69 21. Le Damany P (1904) Un défaut de la hanche humaine : sa double

manifestation, anatomique et physiologique. J Anat Physiol 40:1–21 22. Reikerås O, Hoiseth A (1982) Femoral neck angles in osteoarthri-

tis of the hip, Acta Orthop Scand 53:781–4

23. Kapandji AI (2009) Anatomie fonctionnelle, Tome II. Membre inférieur, 6ème édition. Maloine, Paris

24. Berge C, Orban-Segebarth R, Schmid P (1984) Obstetrical interpretation of the australopithecine pelvic cavity. J Hum Evol 13:573–87

25. Tague RG, Lovejoy CO (1986) The obstetrical pelvis of A.L.

288-1 (Lucy). J Hum Evol 15:237–55

26. Rosenberg KR (1992) The evolution of modern human child- birth, Yearb Phys Anthropol 35:89–124

27. Berge C (1992) Analyse morphométrique et fonctionnelle du pelvis des Australopithèques (Australopithecus afarensis, A. africa- nus) : interprétations locomotrices et obstétricales. In: Toussaint M (ed) Cinq millions d’années, l’aventure humaine–Five million years, the human adventure. E.R.A.U.L. 56, Liège, pp 49-62 28. Häusler M, Schmid P (1995) Comparison of the pelves of Sts

14 and AL 288-1: implications for birth and sexual dimorphism in australopithecines. J Hum Evol 29:363–83

29. Ruff CB (1995) Biomechanics of the hip and birth in early Homo. Am J Phys Anthropol 98:527–74

30. Ruff CB, Trinkaus E, Holliday TW (1997) Body mass and ence- phalization in PleistoceneHomo. Nature 387:173–6

31. Rosenberg KR, Trevathan W (2002) Birth, obstetrics and human evolution. Br J Obstet Gynaecol 109:1199–206

32. Lovejoy CO (2005) The natural history of human gait and posture. Part 1: Spine and pelvis, Gait Posture 21:95–112

33. Lovejoy CO (2005) The natural history of human gait and posture, Part 2: Hip and thigh, Gait Posture 21:113–24

34. Bouhallier J, Berge C (2006) Analyse morphologique et fonctionnelle du pelvis des primates Catarrhiniens : conséquences pour l’obstétrique. C R Palevol 5:551–60

35. Ruff CB (1991) Climate, body size and body shape in hominid evolution, J Hum Evol 21:81–105

36. McHenry HM, Corruccini RS (1978) The femur in early human evolution. Am J Phys Evol 49:473–88

37. Stern JT, Susman RL (1983) The locomotor anatomy ofAustra- lopithecus afarensis. Am J Phys Anthropol 60:279–317 38. Asfaw B (1985) Proximal femur articulation in pliocene homi-

nids. Am J Phys Anthropol 68:535–8

39. Ruff C (1988) Hindlimb articular surface allometry in Hominoi- dea and Macaca, with comparisons to diaphyseal scaling. J Hum Evol 17:687–714

40. MacLatchy LM, Bossert WH (1996) An analysis of the articular surface distribution of the femoral head and acetabulum in anthropoids, with implications for hip function in Miocene hominoids, J Hum Evol 31:425–53

41. MacLatchy LM (1996) Another look at the australopithecine hip.

J Hum Evol 31:455–76

42. Ohman JC, Krochta TJ, Lovejoy CO, et al (1997) Cortical bone distribution in the femoral neck of hominoids: implications for the locomotion ofAustralopithecus afarensis. Am J Phys Anthro- pol 104:117–31

43. Rafferty KL (1998) Structural design of the femoral neck in primates, J Hum Evol 34:361–83

44. Tayton E (2007) Femoral anteversion. A necessary angle or an evolutionary vestige?. J Bone Joint Surg Br 89:1283–8 45. Bonneau N, Libourel PA, Simonis C, et al (2012) A three-

dimensional orientation of the femoral neck orientation. J Anat 221:465–76

46. Perréard Lopreno G, Eades S (2003) Une démarche actualiste en paléoanthropologie : la collection de squelettes de référence. In Besse M, Stahl Gretsch LI, Curdy P (eds) ConstellaSion. Hom- mage à Alain Gallay, Cahiers d’Archéologie romande 95, Lau- sanne, pp 463-72

47. Bonneau N, Bouhallier J, Baylac M, et al (2012) Study of the three-dimensional orientation of the labrum: its relations with the osseous acetabular rim. J Anat 220:504–13

48. Bonneau N, Bouhallier J, Simonis C, et al (2012) Shape variability induced by reassembly of human pelvic bones. Am J Phys Anthropol 148:139–47

49. R Development Core Team (2011) R: A language and environ- ment for statistical computing. R Foundation for Statistical Com- puting, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.

R-project.org/

50. Fabeck L, Farrokh D, Tolley M (1998) Etude de l’antéversion cotyloïdienne par mesure Coxométrique. J Radiol 79:743–50 51. Fabeck L, Farrok D, Descamps PY, et al (1999) Etude du recouv-

rement antérieur de l’acétabulum. J Radiol 80:1636–41 52. Maruyama M, Feinberg JR, Capello WN, D’Antonio JA (2001)

Morphologic features of the acetabulum and femur. Clin Orthop Relat Res 393:52–65.

53. Vandenbussche E, Saffarini M, Delogé N, et al (2007) Hemi- spheric cups do not reproduce acetabular rim morphology. Acta Orthop 78:327–32

(15)

54. Vandenbussche E, Saffarini M, Taillieu F, Mutschler C (2008) The asymmetric profile of the acetabulum. Clin Orthop Relat Res 466:417–23

55. Stern JT, Oxnard CE (1973) Primate locomotion: some links with evolution and morphology, Primatologia 4:1–93

56. Sugardjito J, van Hooff J (1986) Sex-age class differences in positional behavior of Sumatra orangutan (Pongo pygmaeus abe- lii) in Gunung Leuser National Park, Indonesia. Folia Primatol 47:14–25

57. Cant JGH (1987) Positional behavior of female bornean orangutans (Pongo pygmaeus). Am J Primatol 12:71–90

58. Hunt KD (1991) Positional behaviour in the hominoidea. Int J Primatol 12:95–118

59. Thorpe SKS, Holder R, Crompton RH (2009) Orangutans employ unique strategies to control branch flexibility. P Natl Acad Sci 106:12646–51

60. Thorpe SKS, Crompton RK (2005). Locomotor ecology of wild orangutans (Pongo pygmaeus abelli) in the gunung leuser ecosys- tem, Sumatra, Indonesia: a multivariate analysis using loglinear modelling. Am J Phys Anthropol 127:58–78

61. Thorpe SKS, Crompton RH (2006) Orangutan positional behaviour and the nature of arboreal locomotion in hominoidea. Am J Phys Anthropol 131:384–401

62. Baylac M (2010) Rmorph a morphometric library for R. Avail- able from the author: baylac(at)mnhn.fr

63. Gower JC (1975) Generalized Procrustes analysis. Psychometrika 40:33–50

64. Rohlf FJ, Slice DE (1990) Extensions of the Procrustes method fort he optimal superimposition of landmarks. Syst Zool 39:40–59 65. Schultz AH (1953) The relative thickness of the long bones and

the vertebrae in primates. Am J Phys Anthropol 11:277–311

66. McHenry MH (1975) Biomechanical interpretation of the early hominid hip. J Hum Evol 4:343–55

67. Oxnard CE, Yang HC (1981) Beyond biometrics: studies of complex biological patterns, Symp Zool Soc Lond 46:127–67 68. Poss R (1984) Functional adaptation of the human locomotor

system to normal and abnormal loading patterns, Calcif Tiss 36:151–61

69. Corruccini RS, McHenry HM (1978) Relative femoral head size in early Hominids. Am J Phys Anthropol 49:145–8

70. Jungers WL (1988) Relative joint size and hominid locomotor adaptations with implications for the evolution of hominid bipedalism. J Hum Evol 17:247–65

71. Inman VT (1947) Functional aspects of the abductor muscles of the hip. J Bone Joint Surg 29:607–19

72. Pauwels F (1935) Der Schenkelhalsbruch – Ein mechanisches Problem - Grundlagen des Heilungsvorganges. Prognose und kausale Therapie. Z Orthop Chir Suppl 63:38–43

73. McLeish RD, Charnley J (1970) Abduction forces in the one- legged stance, J Biomech 3:191–209

74. Williams GC (1992) Natural selection: domains, levels, and chal- lenges, Oxford University Press, New York

75. Vereecke EE, D’Août K, Aerts P (2006) Locomotor versability in the white-handed gibbon (Hylobates lar): a spatiotemporal analysis of the bipedal tripedal, and quadrupedal gaits. J Hum Evol 50:552–67

76. Prost JH (1967) Bipedalism of man and gibbon compared using estimates of joint motion, Am J Phys Anthropol 26:135–48 77. Rickenmann E (1957) Beiträge zur vergleichenden anatomie

insbesondere des beckens bei Catarrhinen, Basel