Assessing endocranial variations in great apes and humans using 3D data from virtual endocasts.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 145:231–246 (2011) Assessing Endocranial Variations in Great Apes and Humans Using 3D Data From Virtual Endocasts Thibaut Bienvenu,1,2* Franck Guy,1 Walter Coudyzer,3 Emmanuel Gilissen,4,5,6 Georges Roualdès,7 Patrick Vignaud,1 and Michel Brunet1,2 1 Institut International de Paléoprimatologie, Paléontologie Humaine: Evolution et Paléoenvironnements (iPHEP), CNRS/Université de Poitiers, 86022 Poitiers, France 2 Collège de France, Chaire de Paléontologie humaine, 75231 Paris Cedex 05, France 3 Department of Radiology, University Hospitals Leuven, B-3000 Leuven, Belgium 4 Department of African Zoology, Royal Museum for Central Africa, B-3080 Tervuren, Belgium 5 Université Libre de Bruxelles, Laboratory of Histology and Neuropathology, B-1070 Brussels, Belgium 6 Department of Anthropology, University of Arkansas, Fayetteville, AR 72701 7 CHU La Milétrie, Service de Neurochirurgie, 86021 Poitiers, France KEY WORDS brain; endocast; comparative anatomy; evolution; hominoids ABSTRACT Modern humans are characterized by their large, complex, and specialized brain. Human brain evolution can be addressed through direct evidence provided by fossil hominid endocasts (i.e. paleoneurology), or through indirect evidence of extant species comparative neurology. Here we use the second approach, providing an extant comparative framework for hominid paleoneurological studies. We explore endocranial size and shape differences among great apes and humans, as well as between sexes. We virtually extracted 72 endocasts, sampling all extant great ape species and modern humans, and digitized 37 landmarks on each for 3D generalized Procrustes analysis. All species can be differentiated by their endocranial shape. Among great apes, endocranial shapes vary from short (orangutans) to long (gorillas), perhaps in relation to different facial orientations. Endocranial shape differences among African Modern humans are characterized by their large, complex, and specialized brain. Understanding the timing and processes of brain evolution in hominids is an important objective of paleoanthropological studies. We use the term hominid here to denote all taxa that are more closely related to humans (Homo sapiens) than to chimpanzees (Pan troglodytes) and bonobos (Pan paniscus). Human brain evolution can be addressed in two different ways: direct and indirect. The direct approach consists of studying fossil hominid endocranial casts (or endocasts), as these are the only remnants of hominid ancestors’ brains. This direct method is the ‘‘bread and butter’’ of paleoneurology (Falk, 2007). The other way to address human brain evolution is to study and compare brains of our closest extant relatives, the great apes, with those of humans. We use the term apes referring to hominoids excluding humans; great apes are chimpanzees, bonobos, gorillas, and orangutans; African apes are chimpanzees, bonobos, and gorillas. Such comparative neurological data constitute the indirect evidence of human brain evolution. Endocasts constitute the only direct way to access brain morphology when soft tissues are not available, for example in fossil specimens. However, endocasts are only an estimate of brain anatomy, as the brain surface is separated from the internal bony table by the three meningeal layers, containing cerebrospinal ﬂuid and blood vessels. Endocasts, thus, provide additional inforC 2011 V WILEY-LISS, INC. apes are partly allometric. Major endocranial traits distinguishing humans from great apes are endocranial globularity, reﬂecting neurological reorganization, and features linked to structural responses to posture and bipedal locomotion. Human endocasts are also characterized by posterior location of foramina rotunda relative to optic canals, which could be correlated to lesser subnasal prognathism compared to living great apes. Species with larger brains (gorillas and humans) display greater sexual dimorphism in endocranial size, while sexual dimorphism in endocranial shape is restricted to gorillas, differences between males and females being at least partly due to allometry. Our study of endocranial variations in extant great apes and humans provides a new comparative dataset for studies of fossil hominid endocasts. Am J Phys Anthropol 145:231–246, 2011. V 2011 C Wiley-Liss, Inc. mation concerning brain vascularization that isolated brains do not display. In this article, we use the terms brain and endocast interchangeably to refer to our work that deals exclusively with casts of the braincase. Many previous studies concerning fossil hominid endocasts relate to endocranial volume, which provides an estimate of brain size (Conroy et al., 1998, 2000a,b). According to Falk (2007) and Falk et al. (2009), the notion that brain size suddenly began to increase with the evolution of the genus Homo around 2 Ma needs reevaluation. Indeed, Additional Supporting Information may be found in the online version of this article. Grant sponsor: Agence Nationale de la Recherche; Grant number: ANR-09-BLAN-0238; Grant sponsor: Allocation de Recherche of the French Ministère de l’Enseignement Supérieur et de la Recherche. *Correspondence to: Thibaut Bienvenu, iPHEP, 40 Avenue du Recteur Pineau, F-86022 Poitiers Cedex, France. E-mail: firstname.lastname@example.org Received 15 July 2010; accepted 13 December 2010 DOI 10.1002/ajpa.21488 Published online 1 March 2011 in Wiley Online Library (wileyonlinelibrary.com). 232 T. BIENVENU ET AL. with the exclusion of Paranthropus specimens, which may not be directly ancestral to humans (González-José et al., 2008), and the inclusion of ‘‘transitional’’ specimens from Dmanisi (Gabunia et al., 2000; Vekua et al., 2002), brain size appears to increase from before 3 Ma, in gracile australopithecines (Falk, 2007; Falk et al., 2009). Brain size relative to body size (or relative brain size) is often expressed as an encephalization quotient, which is the ratio of actual brain size of any given taxon or specimen on the brain size predicted for a baseline animal of equivalent body size. Encephalization quotients depend on the database used to derive the basic brain/body allometric equations (Holloway et al., 2004a). Encephalization quotients reported below are derived from an ape-based regression. Two important fossil skeletons provide data that enable the assessment of encephalization quotients. The encephalization quotient of Lucy AL 288-1 (Australopithecus afarensis, 3.2 Ma, Ethiopia) is estimated to be one—there are not enough cranial remains of Lucy to measure cranial capacity, but Holloway et al. (2004a) estimate it between 375 and 400 cm3, based on the size of the occipital portion—i.e., its brain size is similar to that predicted for an ape of similar body size. The skeleton from Nariokotome KNM-WT 15000 (Homo erectus, 1.5 Ma, Kenya) has an encephalization quotient of two, while modern humans have encephalization quotients of three. Therefore, relative brain size also increased from 3 Ma until today. The second important issue concerning fossil hominid endocasts relates to their shape, as it can provide information pertaining to the level of neurological reorganization. Shapes of fossil and extant hominid endocasts have been analyzed with traditional morphometrics (Neubauer et al., 2004; Falk et al., 2005; Falk and Clarke, 2007; Falk et al., 2007) and 2D or 3D geometric morphometrics (Bruner et al., 2003; Bruner, 2004, 2008; Bruner and Ripani, 2008; Bastir et al., 2008; Neubauer et al., 2009). Among reorganizational changes that occurred during hominid evolution, the most often cited are 1) expansion of parietal association cortex, concomitant to a reduction of primary visual cortex, resulting in a posteriorly located lunate sulcus, 2) widening of the prefrontal portion of frontal lobes, and 3) development of cerebral asymmetries (Broca’s cap and petalias) (Holloway et al., 2004a; Falk et al., 2009). Much debate has centered on brain size versus neurological reorganization, and which came ﬁrst (Holloway et al., 2001, 2004b). The description of Homo ﬂoresiensis (Brown et al., 2004) and analysis of a virtual endocast of the type specimen LB1 (Falk et al., 2005, 2007) sheds even more light on this ‘‘falsely perceived dichotomy’’ (Gould, 2001). Despite LB1 having an ape-sized brain and an encephalization quotient of 1, its brain shows many features of neurological reorganization (Falk et al., 2009) consistent with the higher cognitive abilities that have been attributed to Homo ﬂoresiensis (Brown et al., 2004; Morwood et al., 2004). Whether reorganization occurred in big- or small-brained ancestors of LB1 remains a matter of debate. However, the ﬁndings associated with LB1 provide signiﬁcant evidence that encephalization cannot be interpreted as an index of some general intelligence (Preuss, 2001). There is a general trend of increasing encephalization and increasing intelligence during hominid evolution, but brain size alone is not a good predictor variable for intelligence. Rather, LB1 extends considerably the realm of possible combinations of brain size and neurological reorganization in fossil hominids (Falk et al., 2009). American Journal of Physical Anthropology Studies aiming to reconstruct human brain evolution through the indirect evidence of comparative neurological data focus on the differences between human and other primate brains (e.g., Rilling and Insel, 1999). Living great apes, especially African apes, have often been viewed as convenient proxies to infer brain morphology of early hominids as they share similar endocranial volumes. Comparative anatomical studies can be performed from cytoarchitectural (Brodmann, 1909; Preuss, 2001) to gross anatomical scales of analysis (Semendeferi et al., 1997; Semendeferi and Damasio, 2000). More recently, functional brain processing has been explored through functional magnetic resonance imaging (fMRI), allowing for the assessment of the relationships between structure and function (e.g., Biswal et al., 2010). Comparative studies have shown that human brains, along with being threefold larger, are not simply allometrically scaled versions of ape brains, and have, for example, an overall larger proportion of neocortex (Rilling, 2006). Contrary to the widespread thought, human brains do not have larger frontal lobes than expected for an ape brain of human size (Semendeferi et al., 1997; Semendeferi and Damasio, 2000). Rather, human frontal lobes are characterized by differences in individual cortical areas and a richer interconnectivity (Semendeferi et al., 2002), demonstrated by an increase in prefrontal white matter volume (Schoenemann et al., 2005). In addition, human brains are more lateralized than those of great apes, reﬂecting hemispheric specializations. Other studies have shown that asymmetries are also present in great ape brains, whether in areas implicated in language in humans (Gannon et al., 1998; Cantalupo and Hopkins, 2001) or in the petalial pattern (relative projection of frontal lobes and occipital lobes, correlated to handedness in humans) (Holloway and Coste-Lareymondie, 1982; Balzeau and Gilissen, 2010). Some studies have concentrated on the description of local shape variations among great apes, such as sulcal (Connolly, 1950) or meningeal arterial pattern (Falk, 1993). However, most studies to date have neglected global shape differences among great ape brains. Holloway (1981) was the ﬁrst to quantitatively address overall endocranial shape variation among extant great apes. He observed global differences for the whole extant sample but did not explore the differences among species. Actual brain anatomy cannot be analyzed in fossil specimens, but endocranial shape provides evidence in addition to endocranial size. A few previous analyses of endocranial shape have used geometric morphometrics. Bruner (2004, 2008) compares the 2D lateral projections of endocasts of extant and fossil humans. Three-dimensional endocranial morphologies have also been analyzed through geometric morphometrics in modern and fossil humans (Bruner et al., 2003; Bruner and Ripani, 2008; Bastir et al., 2008; Neubauer et al., 2009). Before analyzing endocranial shape in fossils, it is important to understand the variation of endocranial shape in living primates. Here we report the ﬁrst global 3D geometric morphometric analysis carried out on endocasts of all extant great ape species and humans. Landmark coordinate analyses are used to quantify and describe the morphological variation among a set of specimens. Patterns of variation in endocranial size and shape among extant great ape species and humans are explored, as well as patterns of variation related to sexual dimorphism. In doing so, this study provides a comparative framework for future analyses of fossil endocasts. 233 HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS MATERIALS AND METHODS Sample The study was conducted on 72 extant cranial specimens, comprising 14 orangutans (Pongo pygmaeus), 16 gorillas (Gorilla gorilla), 16 chimpanzees (Pan troglodytes), 10 bonobos (Pan paniscus), and 16 modern humans (Homo sapiens) (Table 1). All specimens are adults (third molars fully erupted) and males and females are equally represented in each species. Modern human specimens are from Europe (11), Greenland (two), Patagonia (one), Tierra del Fuego (one), and Africa (one). Specimens are from Musée Royal de l’Afrique Centrale of Tervuren (Belgium), Anthropologisches Institut und Museum of Zürich (Switzerland), Natural History Museum of London (United Kingdom), and Staatssammlung für Anthropologie und Paläoanatomie München of Munich (Germany). Data acquisition Each cranium was scanned using a medical computerized tomography (CT) scanner, with a pixel size and a slice thickness adjusted according to specimen cranial size. Pixel size ranged from 0.30 (bonobos) to 0.70 mm (gorillas, humans) and slice thickness from 0.30 (bonobos) to 1.0 mm (humans). Depending on their repository location, specimens were scanned either in the Department of Radiology in Universitair Ziekenhuis (UZ) in Leuven (Belgium), in Kantonsspital in Winterthur (Switzerland), in Hammersmith Hospital in London (United Kingdom), or in the medical practice of Dr. Wuttge and colleagues in Munich (Germany). Data processing Two methods were used to extract virtual endocasts from CT scanned crania (Supp. Info. Figs. S1 and S2). The ﬁrst method is a semi-automated segmentation using Avizo 6.0 software (Visualization Sciences Group). On a coronal CT slice, the endocranial cavity was roughly delineated manually by selecting also a bit of surrounding bone (about 10 pixels). This selection was propagated to adjacent slices using Snakes algorithm without extrapolation (ﬁlter which ﬁts the contour of a selection to the regions with a high contrast) (Supp. Info. Fig. S1). Gray values corresponding to bone were removed from the selection by threshold segmentation on the complete image stack. Manual corrections were made where the endocast did not ﬁt the inner table of TABLE 2. Endocranial volumes (cm3) of ﬁve virtual endocasts produced by both methods TABLE 1. Extant sample Species Males Females Total Species First method Second method % difference Pongo pygmaeus Gorilla gorilla Pan troglodytes Pan paniscus Homo sapiens Total 7 7 8 4 9 35 7 9 8 6 7 37 14 16 16 10 16 72 P. pygmaeus G. gorilla P. troglodytes P. paniscus H. sapiens 464 410 282 352 1285 460 405 278 346 1275 0.9 1.4 1.4 1.7 0.8 Percentages of difference in endocranial volume are indicated. Fig. 1. Right lateral view and ventral view (anterior at top) of an endocast of a female Homo sapiens produced by the second method of extraction (consisting in a segmentation of the bone, followed by a removal of the outer surface and a ﬁlling of the basicranial foramina), gray level-coded to show the surface distance with the endocast produced by the ﬁrst method of segmentation of the endocranial cavity. Gray level map is coded from 0 to 1 mm. The distance between the two surfaces is greatest at foramina because it corresponds to zones either manually segmented on the coronal slices (ﬁrst method), or manually cut in the 3D surface (second method). American Journal of Physical Anthropology 234 T. BIENVENU ET AL. TABLE 3. Landmarks digitized on virtual endocasts of all specimens Number 1, 3, 5, 7, 9 2 4 6 8 10 11 12 13 14 15 16 17, 18 19, 20 21, 23 24, 26, 28, 30, 32, 22 25 27 29 31 33 34, 35 36, 37 Landmark Frontal pole Occipital pole Temporal polea Cerebellar poleb Anterior end of olfactory bulb (cribriform plate) in the sagittal plane 5 foramen caecum Posterior end of olfactory bulb (cribriform plate) in the sagittal plane 5 anterior sphenoid spine Posteriormost point of anterior cranial fossa in the sagittal plane on limbus sphenoidalis Sella Dorsum sellae Endobasion Endopisthion Internal occipital protuberancec 5 torcular herophili Sylvian ﬁssure: point at which the posterior border of the anterior cranial fossa fuses with the endocranial lateral walld Pyramidal root 5 maximum curvature point between transverse and petrous curve Pyramidal apex Endovertex Optic canal, inferior border Foramen rotundum, inferior border Foramen ovale, posterior border Internal acoustic meatus, inferior border Posteriormost point on jugular foramen, i.e., anteriormost point of sigmoid sinus Maximum cerebral width Maximum cerebellar width, sigmoid sinuses included Landmarks 9–16 and 23 are located on the midline, the remainder are bilateral landmarks (right, left). a Temporal pole is deﬁned as anteriormost point on temporal lobe (Falk et al., 2007) and not as point of maximum curvature of the temporal lobe (Bruner, 2004). b Cerebellar pole is the posteriormost point on cerebellar lobe. c If the protuberance is too ﬂat, the landmark is located at the point most representative of the interhemispheric and cerebrocerebellar separation (Bruner and Ripani, 2008). d Bruner and Ripani (2008). bone, for example at foramina or the cribriform plate. Finally, a surface of the endocast was computed from the segmentation data object, yielding a 3D closed object. A second method of virtual extraction was also used. The cranium was automatically segmented by setting a gray value range for the bone on all the coronal slices. A blocking transverse section was set to separate the vault and the basicranium. Gray value range was set more conservative (wider range) for the basicranium than for the vault, as basicranial bone is often very thin and we aimed to avoid many manual corrections. Cavities inside the bone, like paranasal sinuses or mastoid pneumatization, were automatically integrated within the bone. A 3D surface of cranial bone was computed. This surface was then processed by using Geomagic Studio 11 software (Geomagic). The outer surface of the cranium was removed. The remaining inner surface of bone was kept, thus giving a 3D surface of the endocast. Contrary to endocast extraction methods described by Falk et al. (2007) and Neubauer et al. (2009), the internal braincase is closed at foramina after surface extraction of the virtual endocast is performed (Supp. Info. Fig. S2). Foramina are closed near the remaining endocranial surface. Therefore they do not contribute signiﬁcantly to endocranial volume. American Journal of Physical Anthropology Both procedures were compared for ﬁve specimens, one of each sampled extant species. No signiﬁcant differences were observed, either in cranial capacity or in endocranial shape and details (Table 2, Fig. 1). Moreover, no signiﬁcant differences in endocranial volume or shape were found between ﬁve human endocasts extracted by our methods and by that of Neubauer et al. (2009). Differences in cranial capacity between our ﬁrst and second method (always \2% of total volume) are mainly due to differences of segmentation at the level of foramina (Fig. 1). Initially, the ﬁrst method was used. Then it was replaced by the faster second method once we determined that both methods give similar results. Geometric morphometric analysis and statistical analyses Thirty-seven 3D landmarks were identiﬁed and their locations recorded from the endocranial surfaces with Landmark 3.6 software (Wiley et al., 2007) (Table 3, Fig. 2). Endocast orientation may inﬂuence the position of some landmarks and has been the cause of many discussions (Falk, 1986; Holloway and Kimbel, 1986; Holloway and Shapiro, 1992; Falk et al., 1994). Here we start by placing frontal poles and occipital poles, deﬁned respectively as the frontal and occipital ends of the maximal length of each cerebral hemisphere (Connolly, 1950). These poles deﬁne the conventional horizontal plane for orienting endocasts (Holloway and Shapiro, 1992; Falk et al., 1994), and subsequent landmarks (e.g., temporal poles, cerebellar poles, and endovertex) are placed according to this plane, which is equivalent to Frankfurt horizontal plane for skull orientation, but not parallel with it (Falk et al., 1994). To test landmark repeatability, ﬁve specimens (one of each species) were sampled twice. We estimated intraobserver error by subtracting the coordinates of a particular landmark from the coordinates of the same landmark collected during another digitization, following the approach outlined by Corner et al. (1992), since the coordinate system does not change between two repeated measurements of one individual when using CT scans (von Cramon-Taubadel et al., 2007). This enables to calculate localized error to each landmark as the square root of the sum of the squared differences between corresponding coordinates. Error local to each landmark is then averaged across specimens. In addition, Procrustes distances (see below details concerning Procrustes analysis) between repeated measurements of the same individual were compared with Procrustes distances between different individuals, and an unweighted pair group method with arithmetic (UPGMA) mean analysis was performed using the Procrustes distances between specimens. Generalized Procrustes analysis (GPA) was performed on the 3D conﬁgurations using Morphologika 2.5 software (O’Higgins and Jones, 2006), landmark conﬁgurations being scaled to unit centroid size, translated, and rotated, minimizing the sum of squared landmark deviations from the sample mean. Geometric morphometric methods offer many beneﬁts (Rohlf and Slice, 1990; Bookstein, 1991; Dryden and Mardia, 1998), including the possibility to treat size and shape separately and to analyze morphologies as a whole and not only trait by trait. Statistical analyses were performed using Statistica 7.1 software (StatSoft). An analysis of variance (ANOVA) was performed on endocranial volumes to test for differ- 235 HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS Fig. 2. Landmarks used in the Procrustes superimposition (right lateral view and ventral view). Landmarks are displayed on a male Pan paniscus endocast. See Table 3 for numbers correspondence. ences among species and between sexes (acceptable Type I error rate is set at 0.05). While centroid size is an appropriate measure of size in most geometric morphometric analyses, in our application centroid size as computed from 37 landmarks captures only a part of endocranial size. Consequently, we use cranial capacity as a preferred measure of size in our application. Endocranial volume is better suited for comparison with literature. Moreover, endocranial volume and centroid size are highly correlated (R2 5 0.95, P \ 0.05) (Supp. Info. Fig. S3). The Bonferroni post hoc test was used to assess pairwise species differences and intraspeciﬁc sexual differences in endocranial volume. In this test, adjustment is performed by dividing the Type I error rate by the number of tests (Zelditch et al., 2004). A principal component analysis (PCA) based on variance–covariance matrix was performed using Procrustes shape coordinates, yielding new variables (principal components or PCs) that are linear combinations of the original ones and independent of each other. The scree test was used to determine which PCs were signiﬁcant. The scree test, ﬁrst proposed by Cattell (1966), is done by plotting principal components versus their percentage of total variance explained and retaining only components with percentages above a threshold where the decrease of percentages graphically appears to level off. To better describe shape changes associated with each signiﬁcant PC, the inﬂuence of any single landmark on an eigenvector was computed as the square root of the sum of the squared coordinate loadings for that landmark (Baab and McNulty, 2009). To test for endocranial shape differences between taxa or sexes, a multivariate analysis of variance (MANOVA) was performed on signiﬁcant PC scores. Pairwise species differences and species-speciﬁc sex differences were tested along each principal axis, using the Bonferroni post hoc test. Static allometry was studied by regressing (least squares regressions) signiﬁcant PC scores on log endocranial volume and testing for signiﬁcance of squared correlation coefﬁcients. TABLE 4. Mean distances (mm) between corresponding landmarks of ﬁve virtual endocasts digitized twice Landmark 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Average Mean 6 s.d. 2.71 2.32 1.57 1.61 2.04 1.66 1.41 2.28 0.53 0.63 0.58 1.38 0.59 1.57 1.65 2.42 1.99 2.21 2.48 2.06 1.05 1.84 0.94 0.89 1.24 0.56 0.82 0.97 0.80 1.13 1.27 1.68 1.26 1.05 0.41 0.56 0.44 1.37 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 2.50 1.90 1.01 0.54 0.94 1.35 0.89 2.21 0.43 0.36 0.37 1.02 0.66 1.79 1.55 1.64 1.25 1.88 1.96 1.54 0.60 1.71 0.37 0.45 0.67 0.29 0.57 0.49 0.46 0.29 0.23 0.95 0.80 0.98 0.27 0.64 0.18 1.30 Distances greater than 2.00 mm are italicized. See Table 3 and Figure 2 for landmarks identiﬁcation. American Journal of Physical Anthropology 236 T. BIENVENU ET AL. TABLE 5. Endocranial volumes (cm3) Species Sex P. pygmaeus M F Both M F Both M F Both M F Both M F Both G. gorilla P. troglodytes P. paniscus H. sapiens This study, mean 6 s.d. (n) 406.1 345.0 375.6 540.9 451.2 490.4 418.3 366.9 392.6 341.4 345.1 343.6 1399.5 1253.3 1335.6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 45.5 (7) 24.1 (7) 47.2 (14) 41.7 (7) 32.6 (9) 58.1 (16) 29.9 (8) 47.1 (8) 46.4 (16) 37.4 (4) 24.7 (6) 28.4 (10) 128.0 (9) 56.6 (7) 125.0 (16) Isler et al. (2008), mean 6 s.d. (n)a 422.1 338.5 383.5 523.2 459.2 500.5 386.2 350.5 367.6 356.3 326.2 344.3 6 6 6 6 6 6 6 6 6 6 6 6 37.9 29.5 54.0 49.9 32.6 54.0 42.9 30.0 40.7 16.3 36.0 28.7 (57) (49) (106) (40) (22) (62) (55) (60) (115) (6) (4) (10) Holloway (1981), min–max (n) Beals et al. (1984) Falk et al. (2005) Suwa et al. (2009)b 375 422–615 (16) 490 375c 385–449 (15) 412.4 375.9 357.7 344.4 327–439 (16) 1166–1659 (13) 1349 1350 a Calculated from their Appendix 2 listing individual data and considering only one species for Pongo (i.e., clustering Pongo pygmaeus and Pongo abelii in the same species), and one species for Gorilla (i.e., clustering Gorilla beringei and Gorilla gorilla in the same species). b After Fenart and Deblock (1973) and Cramer (1977). c Falk and Clarke (2007) report a mean cranial capacity for adult chimpanzees of 390 cm3 (after Ashton and Spence, 1958). Finally, phenetic afﬁnities derived from endocranial shape data were examined and compared to extant hominoid phylogeny (Goodman et al., 1994; Ruvolo et al., 1994; Ruvolo, 1997). For this purpose, we computed a dendrogram via a UPGMA of Procrustes distances between specimens. RESULTS Intraobserver error The average distance between two corresponding landmarks is 1.37 mm (Table 4). Endocranial poles and notably frontal poles (landmarks 1 and 2), internal occipital protuberance (landmark 16), sylvian ﬁssures (landmarks 17 and 18), and pyramidal roots (landmarks 19 and 20) are more prone to error. Within each species, the Procrustes distance between repeated measurements is inferior to the Procrustes distances between different individuals. In the UPGMA analysis using the Procrustes distances, the resampled individuals cluster together, suggesting that they are more similar to each other than to any other individual in the sample. All these data indicate an acceptable level of error of landmark repeatability and a low within-observer error. Endocranial volume Means and standard deviations of the cranial capacities for each species are listed in Table 5. Our results are consistent with previous literature. Notably, our endocranial volumes are nearly identical to the ones presented in Falk et al. (2005) (Pongo pygmaeus, Gorilla gorilla), Falk and Clarke (2007) (Pan troglodytes), and Isler et al. (2008) (Pan paniscus). The main difference between our estimates of endocranial volumes and those provided in previous studies concerns the mean volume of male bonobos, which is lower than the mean volume of female bonobos in our sample. This difference is probably due to the limited number of individuals (n 5 4 for male bonobos). Differences in endocranial volume between species are statistically signiﬁcant (F[4,62] 5 759.9, P \ 0.05), as well American Journal of Physical Anthropology Fig. 3. Scree plot. Signiﬁcant PCs are PC1-2. as differences between sexes (F[1,62] 5 23.6, P \ 0.05) and species-sex interaction (F[4,62] 5 2.79, P \ 0.05). Bonferroni test shows signiﬁcant differences for each species pairwise comparison, except for Pongo pygmaeus versus Pan troglodytes, Pongo pygmaeus versus Pan paniscus, and Pan troglodytes versus Pan paniscus. Differences between males and females are only signiﬁcant in humans. Dimorphism approaches statistical signiﬁcance in gorillas and a larger sample would probably show signiﬁcant differences in endocranial volume between male and female gorillas. As species-sex interaction is signiﬁcant, species with larger brains (humans and gorillas) show greater dimorphism than species with smaller ones (orangutans, chimpanzees, and bonobos). Endocranial shape In our PCA of endocranial shape, only PC1-2 are signiﬁcant according to the scree test (Fig. 3), representing 55.8% of total variance. Plot of PC1-2 specimen scores are HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS 237 Fig. 4. PC1 (39.5% of variance) versus PC2 (16.3% of variance). Triangles pointing downward represent Pan paniscus, triangles pointing upward Pan troglodytes, circles Gorilla gorilla, diamonds Pongo pygmaeus, and squares Homo sapiens. Solid symbols represent males and open symbols females. Frames represent extreme morphologies. See Figure 2 for frame orientation and landmarks identiﬁcation. Isolated points in lateral view are foramina. presented in Figure 4. PC3 and PC4 are not signiﬁcant but provide additional information for the comparison of species (Supporting Information). Plots of PC1 versus PC3 and PC4 are displayed in Supp. Info. Figure S4. Depending on their species, specimens are differently distributed in PC1-2 shape space (Fig. 4). The ﬁrst principal component mostly separates great apes (lower scores) from humans (higher scores). African apes are also segregated along PC1, from gorillas with lower scores to bonobos with higher scores, while orangutans have PC1 scores which overlap those of gorillas and chimpanzees. Along PC2, gorillas display the lowest scores and orangutans the highest scores, while chimpanzees, bonobos, and humans have intermediate PC scores. The weightings of landmarks on each signiﬁcant PC are presented in Table 6 and are helpful in interpreting the wireframes presented in Figure 4. Specimens scoring higher on PC1 (39.5% of variance) differ from lower scoring individuals in having brains with a more globular shape overall (i.e., they are relatively taller and broader), greater basicranial ﬂexion and shorter basicranium (pattern drawn by foramina compressed anteroposteriorly, but also mediolaterally), more anterior and horizontally oriented foramen magnum, greater posterior projection of the occipital lobes and cerebellar poles more medially positioned (making posterior endocranial morphology more pointed). The olfactory bulb is also longer and more posteriorly positioned, protruding less anteriorly relative to the frontal poles which are located more anteroinferiorly. Pyramids are less obliquely oriented anteriorly and the region between endopisthion and the internal occipital protuberance is more anteroinferiorly angled rather than being more vertical. The second principal component (16.3% of variance) primarily describes the transition from long and narrow to short and wide brains. In specimens scoring higher on PC2, the foramen magnum becomes more oblique through a great elevation of endopisthion (highest loading on PC2, Table 6). The frontal poles and olfactory bulb get further apart. The posterior point of anterior fossa and optic canals become more posteriorly located than foramina rotunda. In PC1-2 shape space, differences in endocranial shape among species and between sexes are statistically signiﬁcant (Wilk’s kspecies effect 5 0.007, F[8, 122] 5 172.4, P \ 0.05; Wilk’s ksex effect 5 0.821, F[2, 61] 5 6.66, P \ 0.05). MANOVA does not reveal a signiﬁcant species-sex interaction in endocranial shape in PC1-2 shape space (Wilk’s ksex*species 5 0.814, F[8, 122] 5 1.65, P 5 0.12). Along PC1, Bonferroni test shows signiﬁcant differences for each species pairwise comparison. Along PC2, differences are signiﬁcant for each species pairwise comparison, except for Pan troglodytes versus Pan paniscus and Pan troglodytes versus Homo sapiens. American Journal of Physical Anthropology 238 T. BIENVENU ET AL. TABLE 6. Loadings of landmarks on each signiﬁcant eigenvector Landmark PC1 PC2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 0.28 0.29 0.15 0.16 0.09 0.10 0.15 0.14 0.11 0.27 0.13 0.13 0.15 0.25 0.36 0.08 0.19 0.18 0.08 0.09 0.09 0.08 0.26 0.11 0.11 0.14 0.13 0.10 0.09 0.11 0.11 0.20 0.19 0.11 0.12 0.08 0.07 0.26 0.25 0.12 0.12 0.13 0.15 0.10 0.11 0.23 0.14 0.21 0.11 0.09 0.08 0.35 0.11 0.12 0.12 0.14 0.15 0.03 0.04 0.11 0.23 0.24 0.08 0.08 0.11 0.13 0.16 0.19 0.13 0.16 0.15 0.21 0.23 0.21 Landmarks with loadings equal or greater than 0.20 are italicized. Comparable loadings of a same landmark on different eigenvectors (e.g., frontal poles) may reﬂect various contributions of x, y, and z coordinates. Differences in endocranial shape are most marked between Homo sapiens and the other species (Fig. 5). These shape differences are notably those associated with PC1, mainly an increase in endocranial globularity and basicranial ﬂexion. Orangutans are characterized by a short braincase, a dorsally oriented foramen magnum, a wide space between frontal poles and olfactory bulb, the latter projecting downward, and widely spaced foramina in the mediolateral direction. Gorillas are characterized by a long and narrow brain (Fig. 5), an extended basicranium, an olfactory bulb protruding anteriorly, an anterior position of optic canals and posterior point of anterior fossa relative to foramina rotunda, and pyramids obliquely oriented anteriorly. The brain of chimpanzees has a low endovertex, posteriorly located maximum cerebellar width, pyramids less tilted anteriorly compared to other great apes, and an olfactory bulb protruding anteriorly like gorillas. Compared to chimpanzees, bonobos are characterized by pyramids more coronally oriented and more inclined anteriorly, a pattern drawn by foramina more compressed anteroposteriorly, a greater basicranial ﬂexion, an olfactory bulb protruding less anteriorly (Fig. 5), American Journal of Physical Anthropology and higher endovertex and maximum cerebral width. In a recent study of MRI scans of hominoid brains, Aldridge (2011) determines cortical and subcortical features that distinguish the human brain from that of apes. Among these, human brains notably show an anteroposterior expansion of the frontal regions, as well as an expansion of the temporal lobes and the anterior parietal regions. Here all of these features are observed endocranially, contributing to overall globularity of the human endocast. Aldridge (2011) also shows patterns of morphology that uniquely differentiate each ape species, among which an anteroposterior elongation of nonfrontal cerebral structures of gorillas that can be observed here in the elongated shape of their endocasts. According to Bonferroni test, sexual dimorphism in endocranial shape is only signiﬁcant along PC1 within gorillas. A separate PCA (not shown) was conducted on mean shapes of male and female gorillas to free from inﬂuences of the variation among species and better assess ‘‘true’’ shape dimorphism within gorillas. Compared to males, endocasts of female gorillas display an overall widening, inferior displacement of the frontal poles, anterior and inferior displacement of endopisthion, anterior displacement of jugular foramina, and posterior displacement of occipital poles. Static allometry Correlations between the ﬁrst two principal component scores and log endocranial volume (Fig. 6) are signiﬁcant for African apes taken together (R2 5 0.58, P \ 0.05, and R2 5 0.40, P \ 0.05, respectively). Hence, endocranial shape differences among African ape species along PC1 and PC2 at least partially reﬂect endocranial size differences, as they all follow a common pattern of allometric shape change. On the contrary, humans are clearly separated from the great apes and orangutans markedly depart from the African apes allometric trajectories (Fig. 6). Correlation between PC1 and log endocranial volume is also signiﬁcant within gorillas (R2 5 0.44, P \ 0.05). Other intraspeciﬁc correlations between PC1-2 and log endocranial volume are not signiﬁcant, but there seems to be a trend of males and females separated within Pongo along PC1 (Fig. 6a), and males of Pan troglodytes seem to display higher endocranial volumes than females. These results suggest that a larger sample may reach statistical signiﬁcance for dimorphism in endocranial volume in Gorilla, Pongo, and Pan troglodytes, and for dimorphism in endocranial shape along PC1 in Pongo. Phenetic afﬁnities Figure 7 shows the result of the UPGMA of Procrustes distances. Overall, individuals from a species cluster together, except for two chimpanzees: one clusters with bonobos, and another is situated at the root of the cluster composed by chimpanzees and bonobos. The tree displays the same topology as extant hominoid molecular phylogeny, when humans are excepted (Goodman et al., 1994; Ruvolo et al., 1994; Ruvolo, 1997). Pongo pygmaeus link at the base of a cluster composed of African apes, both Pan species clustering HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS 239 Fig. 5. Species pairwise Procrustes superimpositions of mean endocasts and corresponding frames. For each comparison, a specimen is put orange and transparent (by order of priority: Homo sapiens, Pongo pygmaeus, Pan paniscus, and Pan troglodytes) while the other one is green and opaque. See Figure 2 for landmarks identiﬁcation. (a) Right lateral view: Upper triangle, endocasts; lower triangle, frames (for readability, foramina are connected). (b) Ventral view: Upper triangle, frames; lower triangle, endocasts. closer to each other than to Gorilla gorilla. Humans link at the root of the cluster formed by great apes, supporting the peculiar morphology of human endo- casts. Therefore, selected landmarks are well-suited to capture endocranial differences among great apes and humans. American Journal of Physical Anthropology 240 T. BIENVENU ET AL. Fig. 6. Log endocranial volume versus PC1 (a) and PC2 (b). Symbols are the same as for Figure 4. Dashed lines are regression axes for African apes. DISCUSSION Interspeciﬁc endocranial shape variation Basicranial ﬂexion and neurocranial globularity. The ﬁrst principal component mainly separates great apes from humans and accounts for well known human traits: neurocranial globularity and basicranial ﬂexion. Neurocranial globularity, or the three-dimensional roundedness of the cranial vault (Lieberman et al., 2003) is reﬂected by dorsal (vertex), anterior (frontal poles), posterior (occipital poles), and lateral (maximum width) expansion. However, this last feature does not appear clearly in humans, except when compared to the relatively narrow gorilla brain (Fig. 5). Neurocranial globularity is linked to the development of visuospatial integration (reorganization of parietal lobes), as well as verbal, cognitive, and social skills (reorganization of frontal lobes) (Bruner, 2004; Holloway et al., 2004a). The other main shape change associated with PC1 is basicranial ﬂexion. According to the spatial packing hypothesis, basicranial ﬂexion is a structural response to an increase in brain size relative to cranial base length (Ross and Ravosa, 1993; Strait, 1999). Cranial base ﬂexion, in combination American Journal of Physical Anthropology with a large brain and a small face, has been shown to contribute to neurocranial globularity in recent human evolution (Lieberman et al., 2002, 2003). Both characters are correlated in the shape changes associated with PC1, stressing that they are not independent from each other. However, the combination of cranial base ﬂexion and neurocranial globularity is driven by the unique morphology of humans. Indeed, a separate PCA excluding humans (not shown) displays an increase in basicranial ﬂexion along PC2—in this analysis of great apes alone, morphological changes observed along PC1 correspond to the ones observed along PC2 when including humans— associated with a lateral expansion, but without the strong dorsal, anterior, and posterior expansions typical of the neurocranial globularity observed when humans are included. Basicranial ﬂexion has also been linked to human cranial modiﬁcations related to posture and bipedal locomotion (Wood Jones, 1917; Weidenreich, 1941), together with other shape changes related with PC1, like shortening of the cranial base, anterior displacement and horizontal positioning of the foramen magnum (Zollikofer et al., 2005), and tilt of the line protuberanceendopisthion (roughly corresponding to the external nuchal plane). Even though the postural hypothesis has been rejected by several studies favoring the spatial packing hypothesis (Ross and Ravosa, 1993; Strait and Ross, 1999; Ross et al., 2004), Lieberman et al. (2000) mention the possibility of an indirect link between basicranial angle and head posture, and McCarthy (2001) suggests that the postural hypothesis may still be relevant for understanding hominid basicranial ﬂexion. The fossil record provides direct evidence that basicranial ﬂexion and neurocranial globularity can exist independently in early hominid evolution. Indeed some fossil hominids are known to have ﬂexed cranial bases without any major increase in brain size or neurocranial globularity compared to great apes (Spoor, 1997; Lieberman et al., 2003). This may reﬂect their bipedality (Spoor, 1997). Bastir et al. (2010) recently conﬁrmed the multifactorial nature of basicranial ﬂexion, showing that only 60% of variance in basicranial ﬂexion among primates is explained by brain and facial size. Endocranial shape and facial morphology. Along PC2, gorillas display the lowest scores, reﬂecting the elongated appearance of their brains, and orangutans the highest scores, corresponding to their short brains. Using both internal and external cranial landmarks, Zollikofer et al. (2008) show correlations between neurocranial shape, size of frontal sinuses and interorbital space, and facial orientation. Their analysis separates orangutans with a shorter neurocranial shape, absence of frontal sinuses, narrow interorbital space, and upwardly oriented (airorhynch) face, from African apes, characterized by a more elongated neurocranial shape, large frontal sinuses, wide interorbital space, and downwardly oriented (klinorhynch) face. The endocranial modiﬁcations we observe along PC2 may have developed in association with the facial modiﬁcations detailed in Zollikofer et al. (2008), but this would need a study of the relationships between endocranium and ectocranium, which is beyond the scope of this article. Middle cranial fossa anatomy and relative disposition of foramina. Bastir et al. (2008) performed a 3D geometric morphometric analysis of the middle cranial fossa in humans, fossil hominids, and chimpanzees. HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS 241 Fig. 7. Phenetic afﬁnities between extant great apes and humans. Specimens were clustered by applying UPGMA on Procrustes distances. Symbols are the same as for Figures 4 and 6. Gray branches represent repeated measurements of the same specimens. They conclude that modern humans differ in having anterolaterally projecting temporal poles relative to optic canals and foramina rotunda (see shape changes associated with PC1 in Fig. 4, and Fig. 8). They also state that in chimpanzees, temporal poles are at the same anteroposterior level as optic canals. Here we ﬁnd that temporal poles are anterior to optic canals in every analyzed species (Fig. 8). This contrasting observation may be due, in part, to the way of digitizing landmarks. Indeed, while we locate our temporal pole at the anteriormost point on the temporal lobe, Bastir et al. (2008) deﬁne their middle cranial fossa pole as the maximal 3D curvature of greater sphenoid wing. Contrary to other studied species, gorillas and humans are characterized by foramina rotunda located posteriorly relative to the optic canals (relatively low PC2 scores) (Fig. 8). In gorillas, this layout seems to be mainly due to an anterior displacement of optic canals relative to the temporal poles (Fig. 8). In humans, compared to great apes, a broad bridge of bone separates the superior orbital ﬁssure from the foramen rotundum (Fig. 9c). This, associated with the laterally extended shape of the ﬁssure, may be considered as a derived condition of robust australopithecines and genus Homo (Rak et al., 1996), depending on the phylogenetic relationships one assumes between these taxa (Strait and Grine, 2004; González-José et al., 2008). Formation of this broad bridge of bone is associated with a more posteriorly placed foramen rotundum, closer to the foramen ovale (Figs. 8 and 9d). This posterior position of the foramen rotundum in humans may be associated with a posteriorly situated distal portion of the maxillary nerve (branch of trigeminal nerve, which passes through the foramen rotundum to innervate mostly the lower face), American Journal of Physical Anthropology 242 T. BIENVENU ET AL. Fig. 8. Ventral view of mean endocasts. Endocasts were placed with the plane deﬁned by frontal poles and occipital poles perpendicular to the view axis. Temporal poles were used to align specimens (straight line). Black points indicate optic canals medially, foramina ovalia posterolaterally, and foramina rotunda. Note the more anterior position of optic canals relative to foramina rotunda in humans and gorillas. Fig. 9. (a) Posterior view of the inside of a chimpanzee cranium showing the relative disposition of optic canals (o.c.), superior orbital ﬁssures (s.o.f.), and foramina rotunda (f.r.). Thick black arrows indicate the thin bridges of bone separating superior orbital ﬁssures from foramina rotunda. d.s., dorsum sellae; m.c.f., middle cranial fossa. (b) Dorsal view of the inside of a chimpanzee cranium. Because of their anterior position, foramina rotunda are not visible. f.o., foramen ovale; s.t., sella turcica. (c) Posterior view of the inside of a human cranium. Thick black arrows indicate the broad bridges of bone separating superior orbital ﬁssures from foramina rotunda. Note also the laterally extended shape of superior orbital ﬁssures. (d) Dorsal view of the inside of a human cranium. Foramina rotunda are visible and much closer to foramina ovalia than in the chimpanzee. perhaps correlated with reduction of the subnasal portion of the face in humans compared to great apes. However, the relationship between subnasal morphology and the point at which the maxillary nerve exits the endocranium deserves further studies. Static allometry among African apes. African apes display common allometric trajectories along PC1 and PC2 (Fig. 6). Except for PC1-related shape differences within gorillas, shape is only correlated with log endocranial volume interspeciﬁcally, not within other species of African apes. Therefore, shape differences associated with PC1 and PC2 among African ape species may be accounted for, at least partly, by size differences. From bonobos with lowest endocranial volumes to gorillas with highest endocranial volumes, these shape differences include a less ﬂexed basicranium, a pattern drawn by American Journal of Physical Anthropology foramina less compressed anteroposteriorly, a more anteriorly protruding olfactory bulb, and a longer and narrower endocast. McNulty (2004) has shown that allometry has a great inﬂuence on hominoid cranial morphology. His phenetic analysis based on ectocranial data corrected for allometry clusters African apes with gibbons, which the author interprets as a conservative hominoid morphology of African apes, especially gorillas. By extrapolating his interpretation to endocranial morphology, observed phenetic relationships between great apes (Fig. 7) might be due partly to a primitive endocranial pattern shared by African apes, rather than to endocranial synapomorphies. According to this hypothesis, the common allometric trajectory shared among African apes would be conservative, with orangutans and humans departing from this primitive trajectory. However, gibbons do not lie along the same HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS 243 allometric trajectory as African apes (unpublished data). Testing which endocranial morphologies are more primitive and which ones are more derived would need further studies, and notably the inclusion of fossil hominoids in the model. ferences. Rather, differences between the brains of men and women appear in the internal organization (proportions of gray and white matter) but are not reﬂected externally. Sexual dimorphism CONCLUSIONS Size sexual dimorphism. Gorillas and humans exhibit the largest endocranial volumes among extant hominoids, and they also show greater differences in endocranial volume between males and females (though dimorphism does not reach signiﬁcance within gorillas). In apes, as in many mammals, sexual dimorphism in body mass is more pronounced in larger species (Rensch’s rule; Isler et al., 2008). Gorillas are the largest living apes, hence the most sexually dimorphic in body mass. As endocranial volume is strongly correlated with body mass among apes (Isler et al., 2008), higher levels of sexual dimorphism in endocranial volume observed in gorillas might be explained by their more pronounced sexual dimorphism in body mass. However, this explanation is not adequate for humans. The greatest part of brain size differences between men and women is not due to differences in body size (Falk, 2001). Moreover, men and women have different neurological ‘‘Bauplans,’’ men having a higher percentage of white matter and women a greater percentage of gray matter (Gur et al., 1999). Falk (2001) suggests a possible evolutionary interpretation of brain differences between men and women. Her hypothesis assumes that discrepancies between the brains of men and women are related to differential reproductive strategies in their ancestors. A higher percentage of gray matter in women reﬂects ancestral selection for skills appropriate for mothering, like anticipation, forward planning, and vocal communication. This may be related to women’s superior performance in language tasks (Falk, 2001). On the other hand, greater proportions of white matter in men are the result of ancestral selection for better performance in visuospatial tasks, allowing for mobility from the natal group, and ﬁnding and competing for mates. Greater proportions of white matter enhance intrahemispheric connections, for example, between posterior visual areas and motor cortex of the frontal lobe. This hypothesis of male brain evolution explains why men perform better at mental rotation of objects, spatial navigation, and guiding or intercepting projectiles (Kimura, 1992). Differences in the proportion of gray to white matter may explain endocranial size differences between men and women. Indeed, smaller endocranial volumes of women require shorter distances for information transfer, hence a lesser need for white matter, which is the myelinated connective tissue for information transfer across distant regions (Gur et al., 1999). Shape sexual dimorphism Endocranial shape dimorphism is signiﬁcant only in gorillas, along PC1. As gorillas show signiﬁcant dimorphism in endocranial volume and PC1 is correlated with log endocranial volume, sexual dimorphism along PC1 is at least partly a consequence of allometry. Even though Bruner and Ripani (2008) found signiﬁcant sexual dimorphism in the human cranial base, it seems that shape sexual dimorphism within Homo, Pan, and Pongo is mostly restricted to the ectocranium (Schaefer et al., 2004). Interestingly, contrary to gorillas, brain size sexual dimorphism in humans is not accompanied by endocranial shape dif- Our analysis of great ape and human endocasts demonstrates that endocranial shape varies greatly among species. Major endocranial traits that distinguish humans from great apes are endocranial globularity, reﬂecting neurological reorganization, and features linked to structural responses to posture and bipedal locomotion, such as basicranial ﬂexion. Among great apes, orangutans are characterized by short brains with a dorsally oriented foramen magnum, and a wide space between frontal poles and the olfactory bulb, the latter projecting downward. Gorillas have long and narrow brains, with the optic canal anteriorly located relative to the foramen rotundum. The brain of chimpanzees displays a low endovertex, a posteriorly located maximum cerebellar width, and pyramids less tilted anteriorly compared to other great apes. The brain of bonobos is characterized by an olfactory bulb projecting less anteriorly compared to that of chimpanzees and gorillas, and a higher maximum cerebral width than in other great apes. Common allometric trajectories for African apes show that endocranial shape differences among these species are partly due to size differences. Aside from endocranial globularity and features linked to bipedality, one peculiar trait of human endocasts is the posterior location of the foramen rotundum relative to the optic canal, and its proximity to the foramen ovale. This trait may be considered as a synapomorphy of robust australopithecines and genus Homo (Rak et al., 1996), depending on the phylogenetic relationships one assumes among these taxa (Strait and Grine, 2004; González-José et al., 2008). Posterior displacement of the foramen rotundum may be associated with a similar displacement of the distal portion of maxillary nerve, which could be correlated to a lesser subnasal prognathism in humans compared to living great apes. Species with larger brains (gorillas and humans) display greater sexual dimorphism in endocranial volume. Whereas sexual dimorphism in brain size in gorillas is at least partly a consequence of sexual dimorphism in body mass, other explanations must be sought for humans. Brain size dimorphism in humans may be due to differential selection in male and female ancestors (Falk, 2001), which would also explain differences in the internal organization of the brain (proportions of gray and white matter). Gorilla gorilla is the only species which displays sexual dimorphism in endocranial shape, differences between males and females being at least partly due to allometry. Brain size dimorphism in humans is not reﬂected in external brain shape. During the past two decades, the development of new techniques such as computed tomography and geometric morphometrics has initiated new perspectives for the study of endocasts, with taxonomic, evolutionary, and functional implications. Application of these techniques here has allowed for investigation of patterns of variation in endocranial shape and size among extant adult great apes and humans. The present analysis constitutes a basis for future work on juveniles and fossil hominids American Journal of Physical Anthropology 244 T. BIENVENU ET AL. (direct evidence of brain evolution) and investigation of the relationship between endo- and ectocranial shape. ACKNOWLEDGMENTS We wish to thank the following institutions and people for allowing us the access to their specimens: Dr. W. Wendelen of the Musée Royal de l’Afrique Centrale (Tervuren, Belgium), Pr. C.P.E. Zollikofer and Dr. M. Ponce de León of the Anthropologisches Institut und Museum (Zürich, Switzerland), the Natural History Museum (London, United Kingdom), and the Staatssammlung für Anthropologie und Paläoanatomie München (Munich, Germany). We also thank the following facilities and people for the CT-scans data acquisition: Pr. Marchal of the Department of Radiology of UZ Leuven (Leuven, Belgium), the Kantonsspital Winterthur (Winterthur, Switzerland), the Hammersmith Hospital (London, United Kingdom), and the medical practice of Dr. Wuttge and colleagues (Munich, Germany). We thank Dr. D. Pushkina, Dr. R. Lebrun, and Dr. F. Bibi for fruitful discussion. We acknowledge all the iPHEP (Université de Poitiers; CNRS: INEE Institut Ecologie et Environnement) members for technical support and administrative guidance. We greatly thank an associate editor and two anonymous reviewers for their helpful comments on earlier drafts of this manuscript. LITERATURE CITED Aldridge K. 2011. Patterns of differences in brain morphology in humans as compared to extant apes. J Hum Evol 60:94–105. Ashton EH, Spence TF. 1958. Age changes in the cranial capacity and foramen magnum of hominoids. Proc Zool Soc Lond 130:169–181. Baab KL, McNulty KP. 2009. Size, shape, and asymmetry in fossil hominins: the status of the LB1 cranium based on 3D morphometric analyses. J Hum Evol 57:608–622. Balzeau A, Gilissen E. 2010. Endocranial shape asymmetries in Pan paniscus, Pan troglodytes and Gorilla gorilla assessed via skull based landmark analysis. J Hum Evol 59:54–69. Bastir M, Rosas A, Lieberman DE, O’Higgins P. 2008. Middle cranial fossa anatomy and the origin of modern humans. Anat Rec 291:130–140. Bastir M, Rosas A, Stringer C, Cuétara JM, Kruszynski R, Weber GW, Ross CF, Ravosa MJ. 2010. Effects of brain and facial size on basicranial form in human and primate evolution. J Hum Evol 58:424–431. Beals KL, Smith CL, Dodd SM. 1984. Brain size, cranial morphology, climate, and time machines. Curr Anthropol 25: 301– 330. Biswal BB, Mennes M, Zuo X-N, Gohel S, Kelly C, Smith SM, Beckmann CF, Adelstein JS, Buckner RL, Colcombe S, Dogonowski A-M, Ernst M, Fair D, Hampson M, Hoptman MJ, Hyde JS, Kiviniemi VJ, Kötter R, Li S-J, Lin C-P, Lowe MJ, Mackay C, Madden DJ, Madsen KH, Margulies DS, Mayberg HS, McMahon K, Monk CS, Mostofsky SH, Nagel BJ, Pekar JJ, Peltier SJ, Petersen SE, Riedl V, Rombouts SARB, Rypma B, Schlaggar BL, Schmidt S, Seidler RD, Siegle GJ, Sorg C, Teng G-J, Veijola J, Villringer A, Walter M, Wang L, Weng X-C, Whitﬁeld-Gabrieli S, Williamson P, Windischberger C, Zang Y-F, Zhang H-Y, Castellanos FX, Milham MP. 2010. Toward discovery science of human brain function. Proc Natl Acad Sci USA 107:4734–4739. Bookstein FL. 1991. Morphometric tools for landmark data: geometry and biology. Cambridge: Cambridge University Press. Brodmann K. 1909. Vergleichende Lokalisationslehre der Großhirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Leipzig: Johann Ambrosius Barth. American Journal of Physical Anthropology Brown P, Sutikna T, Morwood MJ, Soejono RP, Jatmiko, Saptomo EW, Due RA. 2004. A new small-bodied hominin from the Late Pleistocene of Flores, Indonesia. Nature 431:1055–1061. Bruner E. 2004. Geometric morphometrics and paleoneurology: brain shape evolution in the genus Homo. J Hum Evol 47: 279–303. Bruner E. 2008. Comparing endocranial form and shape differences in modern humans and Neandertals: a geometric approach. Paleoanthropology 2008:93–106. Bruner E, Manzi G, Arsuaga JL. 2003. Encephalization and allometric trajectories in the genus Homo: evidence from the Neandertal and modern lineages. Proc Natl Acad Sci USA 100:15335–15340. Bruner E, Ripani M. 2008. A quantitative and descriptive approach to morphological variation of the endocranial base in modern humans. Am J Phys Anthropol 137:30–40. Cantalupo C, Hopkins WD. 2001. Asymmetric Broca’s area in great apes. Nature 414:505. Cattell RB. 1966. The scree test for the number of factors. Multi Behav Res 1:245–276. Connolly CJ. 1950. External morphology of the primate brain. Springﬁeld: Charles C. Thomas. Conroy GC, Falk D, Guyer J, Weber GW, Seidler H, Recheis W. 2000b. Endocranial capacity in Sts 71 (Australopithecus africanus) by three-dimensional computed tomography. Anat Rec 258:391–396. Conroy GC, Weber GW, Seidler H, Recheis W, Nedden DZ, Haile Mariam J. 2000a. Endocranial capacity of the Bodo cranium determined from three-dimensional computed tomography. Am J Phys Anthropol 113:111–118. Conroy GC, Weber GW, Seidler H, Tobias PV, Kane A, Brunsden B. 1998. Endocranial capacity in an early hominid cranium from Sterkfontein, South Africa. Science 280:1730– 1731. Corner BD, Lele S, Richtsmeier JT. 1992. Measuring precision of three-dimensional landmark data. J Quant Anthropol 3: 347–359. Cramer DL. 1977. Craniofacial Morphology of Pan paniscus, Vol. 10 of Contributions to primatology. Basel: Karger. Dryden IL, Mardia KV. 1998. Statistical shape analysis. Chichester: Wiley. Falk D. 1986. Endocast morphology of Hadar hominid AL 16228—Reply. Nature 321:536–537. Falk D. 1993. Meningeal arterial patterns in great apes: implications for hominid vascular evolution. Am J Phys Anthropol 92:81–97. Falk D. 2001. The evolution of sex differences in primate brains. In: Falk D, Gibson KR, editors. Evolutionary anatomy of the primate cerebral cortex. Cambridge: Cambridge University Press. p 98–112. Falk D. 2007. Evolution of the primate brain. In: Henke W, Tattersall I, editors. Handbook of paleoanthropology. Berlin, Heidelberg, New York: Springer-Verlag. p 1133–1162. Falk D, Clarke R. 2007. Brief communication: new reconstruction of the Taung endocast. Am J Phys Anthropol 134:529–534. Falk D, Hildebolt C, Smith K, Morwood MJ, Sutikna T, Brown P, Jatmiko, Saptomo EW, Brunsden B, Prior F. 2005. The brain of LB1, Homo ﬂoresiensis. Science 308:242–245. Falk D, Hildebolt C, Smith K, Morwood MJ, Sutikna T, Jatmiko, Saptomo EW, Imhof H, Seidler H, Prior F. 2007. Brain shape in human microcephalics and Homo ﬂoresiensis. Proc Natl Acad Sci USA 104:2513–2518. Falk D, Hildebolt C, Smith K, Morwood MJ, Sutikna T, Jatmiko, Saptomo EW, Prior F. 2009. LB1’s virtual endocast, microcephaly, and hominin brain evolution. J Hum Evol 57: 597–607. Falk D, Hildebolt C, Vannier M. 1994. Relationship of squamosal suture to asterion on external skull surfaces versus endocasts of pongids: implications for Hadar early hominid AL 162-28. Am J Phys Anthropol 93:435–439. Fenart R, Deblock R. 1973. Pan paniscus et Pan troglodytes. Crâniométrie. Etude comparative et ontogénique selon les méthodes classiques et vestibulaire. Annales du Musée Royal de l’Afrique Centrale 204. HOMINOID ENDOCRANIAL 3D GEOMETRIC MORPHOMETRICS Gabunia L, Vekua A, Lordkipanidze D, Swisher CC III, Ferring R, Justus A, Nioradze M, Tvalchrelidze M, Antón SC, Bosinski G, Jöris O, Lumley M-Ad, Majsuradze G, Mouskhelishvili A. 2000. Earliest Pleistocene hominid cranial remains from Dmanisi, Republic of Georgia: taxonomy, geological setting, and age. Science 288:1019–1025. Gannon PJ, Holloway RL, Broadﬁeld DC, Braun AR. 1998. Asymmetry of chimpanzee planum temporale: humanlike pattern of Wernicke’s brain language area homolog. Science 279:220–222. González-José R, Escapa I, Neves WA, Cúneo R, Pucciarelli HM. 2008. Cladistic analysis of continuous modularized traits provides phylogenetic signals in Homo evolution. Nature 453:775–778. Goodman M, Bailey WJ, Hayasaka K, Stanhope MJ, Slightom J, Czelusniak J. 1994. Molecular evidence on primate phylogeny from DNA-sequences. Am J Phys Anthropol 94:3– 24. Gould SJ. 2001. Size matters and function counts. In: Falk D, Gibson KR, editors. Evolutionary anatomy of the primate cerebral cortex. Cambridge: Cambridge University Press. p xiii–xvii. Gur RC, Turetsky BI, Matsui M, Yan M, Bilker W, Hughett P, Gur RE. 1999. Sex differences in brain gray and white matter in healthy young adults: correlations with cognitive performance. J Neurosci 19:4065–4072. Holloway RL. 1981. Exploring the dorsal surface of hominoid brain endocasts by stereoplotter and discriminant analysis. Philos Trans R Soc Lond B Biol Sci 292:155–166. Holloway RL, Broadﬁeld DC, Yuan MS. 2001. Revisiting australopithecine visual striate cortex: newer data from chimpanzee and human brains suggest it could have been reduced during australopithecine times. In: Falk D, Gibson KR, editors. Evolutionary anatomy of the primate cerebral cortex. Cambridge: Cambridge University Press. p 177–186. Holloway RL, Broadﬁeld DC, Yuan MS. 2004a. The human fossil record. Volume 3: Brain endocasts—the paleoneurological evidence. Hoboken, NJ: Wiley-Liss. Holloway RL, Clarke RJ, Tobias PV. 2004b. Posterior lunate sulcus in Australopithecus africanus: was Dart right? C R Palevol 3:287–293. Holloway RL, Coste-Lareymondie MCdl. 1982. Brain endocast asymmetry in Pongids and Hominids: some preliminary ﬁndings on the paleontology of cerebral dominance. Am J Phys Anthropol 58:101–110. Holloway RL, Kimbel WH. 1986. Endocast morphology of Hadar hominid AL 162-28. Nature 321:536. Holloway RL, Shapiro JS. 1992. Relationship of squamosal suture to asterion in Pongids (Pan): relevance to early hominid brain evolution. Am J Phys Anthropol 89:275–282. Isler K, Kirk EC, Miller JMA, Albrecht GA, Gelvin BR, Martin RD. 2008. Endocranial volumes of primate species: scaling analyses using a comprehensive and reliable data set. J Hum Evol 55:967–978. Kimura D. 1992. Sex differences in the brain. Sci Am 267:119– 125. Lieberman DE, McBratney BM, Krovitz G. 2002. The evolution and development of cranial form in Homo sapiens. Proc Natl Acad Sci USA 99:1134–1139. Lieberman DE, McBratney B, Krovitz GE. 2003. Cranial base ﬂexion and H. Erectus skulls. Science 300:249. Lieberman DE, Ross CF, Ravosa MJ. 2000. The primate cranial base: ontogeny, function, and integration. Yearb Phys Anthropol 43:117–169. McCarthy RC. 2001. Anthropoid cranial base architecture and scaling relationships. J Hum Evol 40:41–66. McNulty KP. 2004. A geometric morphometric assessment of hominoid crania: conservative African apes and their liberal implications. Ann Anat 186:429–433. Morwood MJ, Soejono RP, Roberts RG, Sutikna T, Turney CSM, Westaway KE, Rink WJ, Zhao J-x, Bergh GDvd, Due RA, Hobbs DR, Moore MW, Bird MI, Fiﬁeld LK. 2004. Archaeology and age of a new hominin from Flores in eastern Indonesia. Nature 431:1087–1091. 245 Neubauer S, Gunz P, Hublin J-J. 2009. The pattern of endocranial ontogenetic shape changes in humans. J Anat 215:240–255. Neubauer S, Gunz P, Mitteroecker P, Weber GW. 2004. Threedimensional digital imaging of the partial Australopithecus africanus endocranium MLD 37/38. Can Assoc Radiol J 55: 271–278. O’Higgins P, Jones N. 2006. Tools for statistical shape analysis. Hull York Medical School. Available at: http://hyms.fme.googlepages.com/resources. Preuss TM. 2001. The discovery of cerebral diversity: an unwelcome scientiﬁc revolution. In: Falk D, Gibson KR, editors. Evolutionary anatomy of the primate cerebral cortex. Cambridge: Cambridge University Press. p 138–164. Rak Y, Kimbel WH, Johanson DC. 1996. The crescent of foramina in Australopithecus afarensis and other early hominids. Am J Phys Anthropol 101:93–99. Rilling JK. 2006. Human and nonhuman primate brains: are they allometrically scaled versions of the same design? Evol Anthropol 15:65–77. Rilling JK, Insel TR. 1999. The primate neocortex in comparative perspective using magnetic resonance imaging. J Hum Evol 37:191–223. Rohlf FJ, Slice D. 1990. Extensions of the Procrustes method for the optimal superimposition of landmarks. Syst Zool 39:40–59. Ross CF, Henneberg M, Ravosa MJ, Richard S. 2004. Curvilinear, geometric and phylogenetic modeling of basicranial ﬂexion: is it adaptive, is it constrained? J Hum Evol 46:185– 213. Ross CF, Ravosa MJ. 1993. Basicranial ﬂexion, relative brain size, and facial kyphosis in nonhuman primates. Am J Phys Anthropol 91:305–324. Ruvolo M. 1997. Molecular phylogeny of the hominoids: inferences from multiple independent DNA sequence data sets. Mol Biol Evol 14:248–265. Ruvolo M, Pan D, Zehr S, Goldberg T, Disotell TR, Dornum Mv. 1994. Gene trees and hominoid phylogeny. Proc Natl Acad Sci USA 91:8900–8904. Schaefer K, Mitteroecker P, Gunz P, Bernhard M, Bookstein FL. 2004. Craniofacial sexual dimorphism patterns and allometry among extant hominids. Ann Anat 186:471–478. Schoenemann PT, Sheehan MJ, Glotzer LD. 2005. Prefrontal white matter volume is disproportionately larger in humans than in other primates. Nat Neurosci 8:242–252. Semendeferi K, Damasio H. 2000. The brain and its main anatomical subdivisions in living hominoids using magnetic resonance imaging. J Hum Evol 38:317–332. Semendeferi K, Damasio H, Frank R, Hoesen GWV. 1997. The evolution of the frontal lobes: a volumetric analysis based on three-dimensional reconstructions of magnetic resonance scans of human and ape brains. J Hum Evol 32:375–388. Semendeferi K, Lu A, Schenker N, Damasio H. 2002. Humans and great apes share a large frontal cortex. Nat Neurosci 5:272–276. Spoor F. 1997. Basicranial architecture and relative brain size of Sts 5 (Australopithecus africanus) and other Plio-Pleistocene hominids. S Afr J Sci 93:182–186. Strait DS. 1999. The scaling of basicranial ﬂexion and length. J Hum Evol 37:701–719. Strait DS, Grine FE. 2004. Inferring hominoid and early hominid phylogeny using craniodental characters: the role of fossil taxa. J Hum Evol 47:399–452. Strait DS, Ross CF. 1999. Kinematic data on primate head and neck posture: implications for the evolution of basicranial ﬂexion and an evaluation of registration planes used in paleoanthropology. Am J Phys Anthropol 108:205–222. Suwa G, Asfaw B, Kono RT, Kubo D, Lovejoy CO, White TD. 2009. The Ardipithecus ramidus skull and its implications for hominid origins. Science 326:68e1–68e7. Vekua A, Lordkipanidze D, Rightmire GP, Agusti J, Ferring R, Maisuradze G, Mouskhelishvili A, Nioradze M, Ponce de Leon M, Tappen M, Tvalchrelidze M, Zollikofer C. 2002. A new skull of early Homo from Dmanisi, Georgia. Science 297:85–89. von Cramon-Taubadel N, Frazier BC, Lahr MM. 2007. The problem of assessing landmark error in geometric morphomet- American Journal of Physical Anthropology 246 T. BIENVENU ET AL. rics: theory, methods, and modiﬁcations. Am J Phys Anthropol 134:24–35. Weidenreich F. 1941. The brain and its role in the phylogenetic transformation of the human skull. Trans Am Phil Soc 31:321–442. Wiley DF, Amenta N, Delson E, Rohlf FJ, Hamann B, St. John K, Alcantara DA, Kil YJ, Ghosh D, Gu S, Harcourt-Smith W, Frost S, Rosenberger AL, Tallman L, Disotell T, Motani R, O’Neill R. 2007. Landmark 3.6. Institute for Data Analysis and Visualization (IDAV) group at the University of California, Davis. Wood Jones F. 1917. Arboreal man. London: Arnold. American Journal of Physical Anthropology Zelditch ML, Swiderski DL, Sheets HD, Fink WL. 2004. Geometric morphometrics for biologists: a primer. San Diego: Elsevier Academic Press. Zollikofer CPE, Ponce de León MS, Lieberman DE, Guy F, Pilbeam D, Likius A, Mackaye HT, Vignaud P, Brunet M. 2005. Virtual cranial reconstruction of Sahelanthropus tchadensis. Nature 434:755–759. Zollikofer CPE, Ponce de León MS, Schmitz RW, Stringer CB. 2008. New insights into mid-late Pleistocene fossil hominin paranasal sinus morphology. Anat Rec 291:1506–1516.