Covariation between the cranium and the cervical vertebrae in hominids

Covariation between the cranium and the cervical vertebrae in hominids Mikel Arlegi a , b , * , Ana Pantoja-P erez c , Christine Veschambre-Couture d , Asier G omez-Olivencia e , f , c a Institut Catal a de Paleoecologia Humana i Evoluci o Social (IPHES-CERCA), Zona Educacional 4, Campus Sescelades URV (Edifici W3), 43007 Tarragona, Spain b Universitat Rovira i Virgili, Department d'Hist oria i Hist oria de l'Art, Avinguda de Catalunya 35, 43002 Tarragona, Spain c Centro UCM-ISCIII de Investigaci on sobre Evoluci on y Comportamiento Humanos, Avda. Monforte de Lemos 5 (Pabell on 14), 28029 Madrid, Spain d UMR 5199 PACEA, Universit e de Bordeaux, All ee Geoffroy Saint Hilaire, B^ atiment B8, CS 50023, 33615, Pessac Cedex, France e Departamento de Geología, Facultad de Ciencia y Tecnología, Universidad del País Vasco-Euskal Herriko Unibertsitatea (UPV/EHU), Barrio Sarriena S/n, 48940 Leioa, Spain f Sociedad de Ciencias Aranzadi, Zorroagagaina 11, 20014 Donostia-San Sebasti an, Spain article info Article history: Received 30 June 2020 Accepted 26 October 2021 Available online 8 December 2021 Keywords: Covariation Phylogeny Neck Apes abstract The analysis of patterns of integration is crucial for the reconstruction and understanding of how morphological changes occur in a taxonomic group throughout evolution. These patterns are relatively constant; however, both patterns and the magnitudes of integration may vary across species. These differences may indicate morphological diversification, in some cases related to functional adaptations to the biomechanics of organisms. In this study, we analyze patterns of integration between two functional and developmental structures, the cranium and the cervical spine in hominids, and we quantify the amount of divergence of each anatomical element through phylogeny. We applied these methods to three-dimensional data from 168 adult hominid individuals, summing a total of more than 1000 cervical vertebrae. We found the atlas (C1) and axis (C2) display the lowest covariation with the cranium in hominids (Homo sapiens,Pan troglodytes,Pan paniscus,Gorilla gorilla,Gorilla beringei,Pongo pygmaeus). H. sapiens show a relatively different pattern of craniocervical correlation compared with chimpanzees and gorillas, especially in variables implicated in maintaining the balance of the head. Finally, the atlas and axis show lower magnitude of shape change during evolution than the rest of the cervical vertebrae, especially those located in the middle of the subaxial cervical spine. Overall, results suggest that differences in the pattern of craniocervical correlation between humans and gorillas and chimpanzees could reflect the postural differences between these groups. Also, the stronger craniocervical integration and larger magnitude of shape change during evolution shown by the middle cervical vertebrae suggests that they have been selected to play an active role in maintaining head balance. ©2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction The relationship between form and function is present in many biological structures (e.g., Preuschoft, 2004;Ercoli et al., 2012; Hutchinson, 2012). In human evolution, one of the most explored topics is the study of morphological changes that may have occurred as an adaptation to bipedal locomotion (Robinson, 1972; Bramble and Lieberman, 2004;Sockol et al., 2007;Lovejoy et al., 2009a,b,c;Warrener et al., 2015;Ryan and Sukhdeo, 2016;Ryan et al., 2018). Those studies largely focused on anatomical elements directly related to locomotion, such as the pelvis and lower limbs (Stern, 2000;Pontzer et al., 2009;Grabowski et al., 2011; Grabowski and Roseman, 2015). However, in recent years, the number of studies regarding the vertebral column has increased, with most studies focusing on the lumbar region and giving special attention to differences in the degree of lordosis both between sexes and across hominin species (Whitcome et al., 2007;Been et al., 2012,2014;G omez-Olivencia et al., 2017) and some on the thoracic region (e.g., Bastir et al., 2014,2017;Been et al., 2017; G omez-Olivencia et al., 2018). *Corresponding author. E-mail address: [email protected] (M. Arlegi). Contents lists available at ScienceDirect Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol https://doi.org/10.1016/j.jhevol.2021.103112 0047-2484/©2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Journal of Human Evolution 162 (2022) 103112 The literature regarding the cervical region in primates was, until recently, relatively scarce compared with the other spinal regions (e.g., Schultz, 1942,1961;Slijper, 1946;Francis, 1955a,b; Toerien, 1957,1961;Jenkins, 1969). More recently, the interest in this region has increased, in particular with studies analyzing the morphofunctional interactions with posture and locomotion (Manfreda et al., 2006;Mitteroecker et al., 2007;Been et al., 2014; Nalley and Grider-Potter, 2017;Arlegi et al., 2017,2018;Meyer et al., 2018) and the relationship with the cranium (Nalley and GriderPotter, 2015,2019;Villamil, 2018). These (and other) studies have used several perspectives to evaluate this functional relationship not only in primates but also in other mammal groups: for example, approaches based on the biomechanical analysis of the cranium (Demes, 1985); kinematic analyses of this complex in the wild and in captivity (Bramble, 1989;Strait and Ross, 1999;Dunbar and Badam, 2000;Cromwell et al., 2001;Choi et al., 2003;Dunbar et al., 2008;Zubair et al., 2019); analyses from radiographs, photographs, electronic sensors, and dissections (Vidal et al., 1986;Graf et al., 1995a,b;Benoit et al., 2020;Jorissen et al., 2020); approaches based on the morphological correlation and integration among traits (Nalley and Grider-Potter 2015;Villamil, 2018); and musculoskeletal analyses to define modules in the craniocervical complex (Diogo et al., 2008,2017;Diogo and Wood, 2011;Esteve-Altava et al., 2015;Arnold et al., 2017a;Powell et al., 2018;Boyle et al., 2020). Broadly, these studies revealed that despite postural differences in mammals, no substantial differences exist among taxa in maintaining cranial balance against gravity at rest (Vidal et al.,1986;Graf et al.,1995a,b). In this passive posture, the mammal neck adopts an s-shaped vertical position to minimize the distance between the mass of the head and the weight-bearing cervicothoracic junction, thus reducing dorsal neck muscular stress (Vidal et al., 1986;Graf and Wilson, 1989). In quadrupeds, the vertical position of the neck requires high dorsiflexion of the cervicothoracic articulation (C6eT2) and hyperextension of the atlanto-occipital joint, which also allows the adjustment of head orientation and gaze (Vidal et al., 1986;White and Panjabi, 1990;Graf et al., 1995a;Nalley and Grider-Potter, 2019). The important functional role played by the cranial and caudal cervical modules contrasts with the midcervical module, which is mainly circumscribed to axial rotation and does not show functional specializations (Graf et al., 1995a,b; Arnold, 2020). However, species that do not display a complete quadrupedal posture, such as some primates, and especially modern humans, show a more limited range of motion in the atlantooccipital articulation; thus, they circumscribe most of the craniocervical motions in the midsagittal plane to the cervicothoracic articulation (White and Panjabi, 1990;Graf et al., 1995a,b). The relatively slight differences observed at rest between quadrupedal and nonquadrupedal mammals increase during locomotion. The former reorients the neck horizontally during exertion, whereas humans and plausibly other upright mammals do not (Vidal et al., 1986;Graf et al.,1995a,b;Dunbar and Badam,1998;Strait and Ross, 1999). The concepts of integration and modularity refer to the degree of interaction between the characters of one or more anatomical structures (Olson and Miller, 1958). Both concepts have been defined as important in the phenotypic evolution of organisms from a developmental, genetic, and/or functional point of view (e.g., Olson and Miller, 1958;Cheverud, 1996;Wagner, 1996; Goswami et al., 2014). Morphological integration describes high degree of correlation within subsets of morphological traits, which may result in long-term coevolution (Cheverud, 1996). Modularity refers to the relative independence of traits that are part of different developmental or functional regions. In evolutionary studies, these concepts are crucial for the reconstruction and understanding of how morphological changes occur in organisms because they can facilitate or restrict the evolution of their characters in specific directions (Wagner, 1996;Hallgrímsson et al., 2007;Goswami and Polly, 2010;G omez-Robles and Polly, 2012). Because integration can enhance or constraint morphological evolution, establishing how patterns of integration have evolved concurrently with the morphology can help determine the evolution of a group (Wagner, 1988;Grabowski et al., 2011). In general terms, it has been proposed that integration patterns are relatively constant in species (Goswami, 2006;Porto et al., 2009;Bardua et al., 2019;Watanabe et al., 2019). However, both the patterns and the magnitudes of integration may vary across species (Marroig and Cheverud, 2001;Marroig et al., 2009;Porto et al., 2009; Goswami and Polly, 2010). Detecting potential differences in integration patterns is critical because these changes may indicate morphological diversification as a result of possible adaptation to selection pressures affecting evolutionary trajectories, likely related to functional effects of the mechanics of organisms (Wagner and Schwenk, 2000). The number of vertebrae in the vertebral column is regulated by the expression of the Hox genes, and those of the paralog groups 4 and 5 control the organization of the cervical region (Kessel and Gruss, 1991;Burke et al., 1995;Galis, 1999a;Wellik and Capecchi, 2003). In mammals, almost all species present a fixed number of seven cervical vertebrae (Bateson, 1894;Johnson and O'Higgins, 1996;Galis, 1999b;Narita and Kuratani, 2005;Varela-Lasheras et al., 2011;Buchholtz, 2014;B€ ohmer, 2017;B€ ohmer et al., 2018), which, at the same time, are internally organized into three functional and developmental modules: upper (C1eC2), middle (C3eC5), and lower cervical (C6eC7; Arnold et al., 2016;Randau et al., 2017). Other studies proposed a slightly different subdivision of the cervical spine, which includes either the cranial base (CB) as part of the upper module (i.e., CBeC1) or the cranium and thoracic spine for some species (Arnold et al., 2017a;Villamil, 2018). The cervical spine is a transitional functional region between the head and the rest of the vertebral column, where each cervical module plays a different functional role (Graf et al., 1995b;Arnold, 2020). In the synapsid/mammal transition, these functional modules did not evolve at the same time; they started with the appearance of the upper cervical module (an early atlas-dens-axis joint) and finished by the consolidation of a lower cervical module (Buchholtz et al., 2012;Arnold, 2020 and references therein). Regarding the morphological link between the cranium and the cervical region in primates, some researchers have found certain correlations between specific characters, which could have implications for posture and locomotor behaviors (Strait and Ross, 1999; Nalley and Grider-Potter, 2017;Villamil, 2018). In general, taxa with more horizontal necks show characteristic cervical traits to avoid vertebral articular displacement (e.g., coronal orientation of the articular facts) and for the insertion of the large epaxial musculature that supports the long load arm that results from this posture (e.g., larger transverse and spinous processes; Adams and Moore, 1975;Ebraheim et al., 2008;Nalley and Grider-Potter, 2015; Arlegi et al., 2017). Villamil (2018) found that the CB and the cervical vertebrae in hominoids, especially the C1 and the central cervical vertebrae (C3eC5), are strongly integrated. However, it was concluded that body posture and locomotion are relatively weak selection pressures in the morphology of cervical vertebrae. Studies using anatomical network analysis also searched for functional, evolutionary, and/or developmental modules in the head-neck by combining the analyses of hard (bone/cartilage) and soft tissues (muscles; Diogo et al., 2008,2017;Diogo and Wood, 2011;Esteve-Altava et al., 2015;Powell et al., 2018;Boyle et al., 2020). These show that the head and neck muscles in primates, compared with other anatomical regions, are a better match for the M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 2 most recent molecular tree (Diogo and Wood, 2011;Boyle et al., 2020) and that this complex is mainly defined by function (Esteve-Altava et al., 2015). They further stress that modern humans have more head and neck muscles than any other primate (Powell et al., 2018), and, interestingly, this leads to the least complex and most derived musculoskeletal system in this region among hominoids (Diogo and Wood, 2011,2017;Powell et al., 2018). The main objective of this study is to quantify and analyze patterns of correlation and the magnitudes of integration between the morphology of the entire cranium and the cervical vertebrae in hominids. Moreover, factors that influence integration on structures in the evolutionary process, such as the effect of size and phylogeny, will be considered. Here, we test the following hypotheses: 1) The morphological differences among functional and developmental cervical modules will be reflected in the degree of craniocervical integration (e.g., Villamil, 2018). 2) Integration is expected to be stronger in species with larger musculoskeletal features in the dorsal neck, which would indicate greater mechanical bending loads in the muscles responsible for counterbalancing the gravitational forces acting on the head (Adams and Moore, 1975). 3) Despite patterns of integration being relatively constant in mammals, we expect certain differences in the pattern of craniocervical correlation among groups that show different postural and locomotor repertoires (Grabowski et al., 2011; Grabowski and Roseman, 2015). 4) The cervical vertebrae in hominids present certain evolutionary disparities (e.g., more directional or stabilizing selection), and these differences would correspond with the internal modular division of the cervical spine (Arnold, 2020 and references therein). 5) Highly integrated sets of traits are mainly aligned along the axis of size-related variation (Marroig et al., 2009;Porto et al., 2009); therefore, removing the effect of size will result in a reduction in the levels of integration (e.g., Porto et al., 2013;Arlegi et al., 2018). 2. Materials and methods The sample studied in this work comprises the cranium and the seven cervical vertebrae of 160 adult individuals of the family Hominidae, totaling 160 crania and 1071 cervical vertebrae: 43 Homo sapiens,46Pan troglodytes,12Pan paniscus,45Gorilla gorilla, 8Gorilla beringei graueri,3Gorilla beringei, and 3 Pongo pygmaeus (Supplementary Online Material [SOM] Table S1). The surface of each anatomical element was scanned in two views, cranial and caudal, using a Go!SCAN 20 (with a resolution of 0.1 mm for the vertebrae and 0.4 mm for the crania) and later virtually assembled into a single three-dimensional (3D) object using VXelements software v. 6.3 (Creaform Inc., L evis). Both geometric morphometric (GM) and traditional morphometric (TM) methods were used to analyze covariation between the cranium and the cervical vertebrae in Hominidae. GM methods were used to perform the analyses at the interspecific level, and TM methods were used at the intraspecific level. Before performing the analyses, landmarks that could not be captured due to damage in the bone were, if possible, estimated using bilateral symmetry and otherwise calculated using partial least squares (PLS; Bookstein et al., 1990;Rohlf and Corti, 2000). Overall, less than 2% of the total landmarks were estimated. All statistical analyses were performed in R v. 4.0.2 (R Core Team, 2020), and, more specifically, for GMs using the package ‘geomorph’v. 3.2.1 (Adams et al., 2020). 2.1. Data collection Three-dimensional landmarks representing the morphology of each anatomical element were virtually captured from the created 3D scan models: 33 landmarks in the cranium, 27 in the atlas (C1), 33 in the axis (C2), and 34 in the subaxial cervical vertebrae (C3eC7; see SOM Tables S2eS5 for landmark definitions; Figs.1 and 2) using Viewbox 4 software v. 4.5.0 (dHAL software, Kifissia). A high percentage of C3eC5 vertebrae of H. sapiens present with bitubercularity of the tip of the spinous process (G omez-Olivencia et al., 2013). In these cases, the landmarks corresponding to the most dorsal point of the spinous process were placed virtually in the midsagittal line. Before performing the analyses, those landmarks that could not be captured owing to damage in the bone were estimated using bilateral symmetry if possible and otherwise were calculated using PLS regression (Bookstein et al., 1990;Rohlf and Corti, 2000). Overall, less than 2% of the total landmarks were estimated. Owing to the difficulty of obtaining individuals in the collections that included both the cranium and all seven cervical vertebrae, those that had the cranium and a minimum of five out of the seven cervical vertebrae were selected. Of the studied individuals, 88% presented the eight anatomical elements examined here (i.e., cranium and seven cervical vertebrae) and the remaining 12% (n¼20) only lack one or two anatomical elements. 2.2. Geometric morphometrics All analyses using GM techniques were performed at the interspecific level including all taxa. Before performing the statistical analyses, we conducted a generalized Procrustes analysis (GPA; Rohlf and Slice, 1990) from the raw 3D coordinates of each cervical vertebra and the cranium separately including all taxa to remove the information related to size, position, and orientation. Integration and pairwise comparisons We quantified the degree of morphological integration between shape variables that describe the different elements of the cranium and cervical vertebrae. We used the two-block PLS regression method, using the ‘integration.test’function of the package ‘geomorph’(Adams et al., 2020). This method uses the decomposition of the between-block covariance matrix and searches for pairs of new axes that represent the maximum amount of covariance between the two blocks. As each axis of each block (e.g., axis 1 of block 1) only correlates with the corresponding axis of the other block (e.g., axis 1 of block 2), the covariance pattern can only be analyzed by a pair of PLS axes at a time (Bookstein et al., 1990; Klingenberg and Zaklan, 2000;Rohlf and Corti, 2000). The results of this procedure are appropriate because it yields values that are unaffected by sample size or by the number of variables (Adams and Collyer, 2016). The significance was calculated by comparing the obtained PLS correlation (r PLS ) values with those resulting from a random permutation of the individuals in one partition relative to those in the other (Bookstein et al., 2003;Adams and Collyer, 2016). Then, to test whether each cervical vertebra presented a significantly different magnitude of integration with the cranium compared with the other vertebrae, we calculated the effect sizes of each pairwise PLS analysis using the ‘compare.pls’function of ‘geomorph’(Collyer et al., 2015;Adams and Collyer, 2016). This analysis, rather than using the PLS correlation coefficient, performs two-sample Z-tests, which are robust to differences in sample size and the number of landmarks. Phylogenetic signal Phylogenetically related species share an evolutionary history and thus tend to display similar trait values due to their common ancestry (Felsenstein, 1985). To evaluate shape divergence over the evolution of the cranium and the M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 3 cervical vertebrae, we calculated the phylogenetic signal of our data using the phylogenetic information of hominids from the 10kTrees Project platform (Arnold et al., 2010) and the Procrustes mean average of each species for each element independently. To estimate the phylogenetic signal, we used the function ‘physignal’ of ‘geomorph,’which is based on the multivariate version of the K-statistic (K mult :Adams, 2014). This method estimates the degree of phylogenetic signal in a data set relative to what is expected under a Brownian motion of evolution. Conversely to previous methods, K mult is appropriate for highly dimensional multivariate data avoiding type I errors. Significance was estimated via shape data permutation among the tips of the phylogeny (Adams, 2014; Adams and Collyer, 2019). Finally, we explored patterns of cranial and cervical vertebrae shape evolution by projecting the phylogeny onto the morphological morphospace represented by the first two principal components (PC1 and PC2). Phylogenetic PLS analysis To account for relatedness among the hominid species in our sample, we calculated craniocervical integration while accounting for the phylogenetic relationships among taxa using phylogenetic PLS analyses. This approach, implemented in the function ‘phylo.integration’of ‘geomorph,’also displays appropriate type I error rates and constant levels of integration irrespective of the number of species or trait dimensions (Adams and Felice, 2014). The significance of the analyses was performed via permutation analysis in the same way as for the PLS analysis (discussed earlier). 2.3. Traditional morphometry All analyses from the Traditional morphometry section were performed at the intraspecific level, that is, including only the three species with the greatest sample sizes (i.e., H. sapiens,G. gorilla, and P. troglodytes). Here, we analyzed covariation between the cranium and the cervical vertebrae using linear measurements from both raw data sets and size-adjusted (removing the influence of size). First, linear measurements were derived from the 3D coordinates using the ‘interlmkdist’function of ‘geomorph.’These linear variables were selected from a series of standard measurements that best represent the morphology of the cranium and each cervical vertebra (SOM Tables S6 and S7;McCown and Keith, 1939;Martin and Saller, 1957;Howells, 1973;Br€ auer, 1988). Six variables were selected for the atlas (C1), nine for the axis (C2), and another nine for each of the C3eC7 vertebrae (Fig. 3). Correlation analyses between two elements require that they both have the same number of variables, so the nine variables selected for the cranium were reduced to six for the analyses with the atlas (SOM Table S8). After that, to test whether sex was a significant source of variation in the data sets that needed to be removed, we accounted for the relative amount of shape variation attributable to sex on each vertebral and cranial element per species by creating a linear model and estimating the probability via analysis of variance. All analyses resulted in nonsignificant differences between sexes, and thus, we did not correct sex variation from the data sets. PLS integration analysis With the obtained linear variables from the raw 3D coordinates, we first analyzed the magnitudes of integration between the cranium and the cervical vertebrae in the three species using the ‘integration.test’function of ‘geomorph,’the same method applied for GMs (discussed earlier). Between variables pairwise correlation test Next, we analyzed patterns of correlation between the cranium and the cervical vertebrae in the three species by quantifying the correlation between pairs of variables. To do so, we calculated the Pearson correlation coefficient and significance between all the cranial variables, on the one side, and all the variables of each vertebra, on the other side, using the ‘cor.test’function of the R package ‘stats’v. 4.0.2 (R Core Team, 2020). Figure 1. Landmarks (red dots) and linear measurements (blue lines) taken of the cranium in this study. Landmarks: upper left, cranium in left lateral view; lower left, in caudal view; upper center, in ventral view. Linear measurements: lower center, in left lateral view; lower center, in ventral view. Note that not all landmarks are visible in all views. See SOM Tables S2 and S6 for landmark and linear measurements definition. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 4 Influence of size Next, we assessed the potential influence of size on integration. To do so, we calculated the geometric mean of each anatomical element (cranium and cervical vertebrae) using their linear measurement values (six for the atlas and nine for each of the other elements), and we used it as a proxy for size (Darroch and Mosimann, 1985;Jungers et al., 1995). To obtain ‘size-corrected’ data sets, we divided the raw values of the linear measurements by the geometric mean for each vertebra and the cranium (Coleman, 2008;Pablos et al., 2013;Arlegi et al., 2017), hereafter size-adjusted. Then, we repeated the two previous analyses (i.e., ‘integration.test’and ‘cor.test’) using the obtained size-adjusted data sets. In addition, we analyzed the amount of cervical shape variation explained by cranial size (i.e., allometry) and the differences in the allometric pattern among species. To do so, we performed a regression analysis using the linear measurements from the raw data sets as dependent variables and the cranial size represented by the geometric mean as the independent variable using the ‘procD.lm’function of ‘geomorph’and the ‘pairwise’function of ‘RRPP’v. 0.6.2 (Collyer and Adams, 2018, 2021) to calculate the angle between the male and female regression vectors and its significance. 2.4. Repeatability of the data sets Finally, to ascertain the reliability of our results, and following Melo et al. (2016), we tested the repeatability of the raw data sets (e.g., C3 G. gorilla,C4P. troglodytes) by bootstrapping each data set 10,000 times and comparing the original and the obtained covariances matrix using random skewers analysis (mean ¼0.880, median ¼0.881; SOM Table S9). 3. Results 3.1. Geometric morphometrics Interspecific integration and pairwise comparisons All analyses yielded high and significant results (Table 1), with craniocervical magnitudes of integration ranging between r PLS ¼0.503 (cranium/atlas) and r PLS ¼0.832 (cranium/C6). The atlas and axis revealed the lowest values of craniocervical covariation followed by the C7 and C3. Thus, central cervical vertebrae (i.e., C4eC6) showed higher values of covariation with the cranium than those located more peripherally in the cervical spine. The altas and axis yielded significantly different values of magnitudes of craniocervical integration compared with the rest of the vertebrae (Table 2). The rest of the pairwise comparison did not reveal significant differences among them; however, it is worth remarking that the lowest differences were obtained among middle cervical vertebrae (i.e., C4eC6). These results relatively support our hypothesis that the degree of craniocervical integration would be different between vertebrae from different cervical modules. Figure 2. Landmarks used in this study in the cervical vertebrae. Note that not all landmarks are visible in all views. See SOM Table S3 for landmark definition. The figured vertebrae belong to Pan troglodytes. M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 5 Phylogenetic signal The results of the phylogenetic analyses are shown in Table 3. They revealed that only the cranium, C3, and C4 vertebrae shapes exhibited significant (though weak) phylogenetic signal, indicating that cranial and cervical vertebrae shapes from these taxa resemble each other less than expected under the Brownian motion model of evolution. The atlas and axis exhibited the lowest values (K mult ¼0.052 and 0.058, respectively) and the cranium the highest (K mult ¼0.337), indicating lower and higher magnitudes of shape change during evolution, respectively. The subaxial cervical vertebrae displayed a trend that increases from C3 to C4 (maximum value, K mult ¼0.296) and then decreases toward C7 (minimum value, K mult ¼0.112). This supports our hypothesis that cervical vertebrae in hominids present a divergence in the magnitudes of evolutionary variation. Phylomorphospaces Visual information of the phylomorphospaces is shown in Figure 4. If the magnitudes of phylogenetic signal (discussed earlier) indicates the magnitude and the direction of shape divergence in the process of evolution (Klingenberg and Gidaszewski, 2010), the phylomorphospaces allow one to visualize the history of morphological diversification of a clade (Sidlauskas, 2008). In Figure 4, we can observe that the cranium and the subaxial cervical vertebrae show a similar evolutionary Figure 3. Linear measurements of the cervical vertebrae used in this study. The numbers indicate the landmarks from which the linear measurements have been calculated. Table 1 Integration values (r PLS ), effect sizes, and standard errors (SE) between the cranium (Cr) and the cervical vertebrae (C1eC7) at the interspecific level (Hominidae). a r PLS pEffect size SE Cr/C1 0.503 <0.001 4.026 0.004 Cr/C2 0.594 <0.001 5.874 0.003 Cr/C3 0.772 <0.001 7.200 0.003 Cr/C4 0.825 <0.001 7.587 0.004 Cr/C5 0.776 <0.001 6.014 0.004 Cr/C6 0.832 <0.001 7.506 0.004 Cr/C7 0.707 <0.001 5.544 0.004 a Significant values are indicated in bold (p-value <0.05). Table 2 Pairwise differences in PLS effect sizes comparing levels of craniocervical morphological integration in hominids. a Cr/C1 Cr/C2 Cr/C3 Cr/C4 Cr/C5 Cr/C6 Cr/C7 Cr/C1 0.062 <0.001 <0.001 <0.001 <0.001 <0.001 Cr/C2 1.867 0.003 <0.001 0.001 <0.001 0.059 Cr/C3 4.697 2.926 0.314 0.625 0.359 0.357 Cr/C4 5.522 3.832 1.006 0.602 0.928 0.063 Cr/C5 5.113 3.376 0.488 0.521 0.666 0.168 Cr/C6 5.455 3.756 0.917 0.090 0.432 0.075 Cr/C7 3.632 1.891 0.921 1.861 1.380 1.779 a Significant values are indicated in bold (p-value <0.05). Values are represented in the lower diagonal, and p-values in the upper diagonal. Table 3 Phylogenetic signal of the cranium and cervical vertebrae in hominids represented in K mult values. Phylogenetic signal pEffect size Cranium 0.337 0.033 1.667 C1 0.052 0.479 0.043 C2 0.058 0.239 0.557 C3 0.186 0.028 1.757 C4 0.296 0.020 1.691 C5 0.283 0.064 1.796 C6 0.226 0.063 1.762 C7 0.112 0.085 1.323 Significant values are indicated in bold (p-value <0.05). M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 6 trend. Hominine groups are separated along, and almost in parallel, the axis of the PC1, showing a strong divergence in opposite directions for gorillas and humans. The genus Pongo diverges from the rest of the family, mostly in the direction of the PC2. In the atlas, P. paniscus and H. sapiens on the one hand, and P. troglodytes and gorillas on the other hand, show shorter distances between them than with respect to their common ancestor, indicating a certain homoplasy in shape. Finally, it is difficult to ascertain the evolutionary ratio based only on the phylomorphospaces; however, except for the atlas, the longer branches of humans indicate faster divergence in this species than those of the other taxa. Phylogenetic PLS The results of the phylogenetic PLS analyses are shown in SOM Table S10. All pairwise tests yielded high and significant values (r PLS >0.965). In contrast to the phylogenetic uncorrected analyses, none of the vertebrae showed significant differences in the r PLS values compared with the rest of the vertebrae. 3.2. Traditional morphometric analyses of raw data This section presents the integration analyses between the cranium and the cervical vertebrae based on linear variables at the intraspecific level. For this purpose, only the three species with larger sample sizes (H. sapiens,P. troglodytes, and G. gorilla) were assessed. PLS integration analysis The results of the PLS analysis from raw data are presented in Table 4 and Figure 5. They show that all cervical vertebrae, except for the atlas in humans and chimpanzees, were significantly integrated with the cranium for the three species. The three groups showed lower values in the atlas, which significantly differ from the rest of the vertebrae, except for the C3 in gorillas (SOM Table S11). In general terms, gorillas possessed the highest values of covariation, followed by chimpanzees and humans. Overall, the pattern of changes in the magnitudes of covariation moving down the vertebral column was similar in different species, but with a different magnitude of Figure 4. Phylomorphospaces of the cranium and the seven cervical vertebrae in Hominidae. The phylogenetic tree of this family was superimposed onto the first two principal components of each anatomical element. The red dot indicates the root of the tree, and the numbers on the branches (internal nodes) represent the last common ancestors for the respective pairs of lineages. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 7 integration, except for the C3 in humans, which departed the pattern followed by chimpanzees and gorillas (Fig. 5). Pairwise comparisons among the three species revealed that gorillas significantly differ from humans and chimpanzees in the craniocervical magnitude of integration in the seven cervical vertebrae (SOM Table S12). In contrast, humans and chimpanzees only displayed significant differences in those vertebrae located in the midlower cervical spine (i.e., C4, C5, and C6). These results support the hypothesis that species with larger musculoskeletal features in the dorsal neck display a higher degree of integration. Between variables pairwise correlation test The results revealed that the correlation between the cranium and the cervical vertebrae in the three species occurred at different degrees and among different traits (Table 5 and Fig. 6;SOM Table S13). In general terms, in gorillas and chimpanzees, the highest craniocervical correlations were shown between either the length of the CB or the length of the face in the cranium and the maximum dorsoventral or transversal diameter traits in the cervical vertebrae (Table 5). In humans, in contrast, a clear correlation pattern did not exist between craniocervical variables. Indeed, in this taxon, the stronger correlation values were displayed in traits related to maximum cranial length and vertebral body height, transverse, and dorsoventral diameter (i.e., M1, M5, and M8). Also, we can observe how, besides magnitude of integration (higher values in gorillas), gorillas and chimpanzees present a more similar pattern of correlation among craniocervical traits throughout the cervical spine compared with humans. They show the highest correlations in the upper left of each diagram, where variables related to the midsagittal plane are located (Fig. 6). Conversely, in humans, the distribution of the highest correlations is not concentrated in that area of the diagram and shows a more heterogeneous pattern throughout the cervical region. These Table 4 Integration values (r PLS ), effect sizes (E-size), and standard errors (SE) from the covariation analysis between the cranium and the cervical vertebrae at the intraspecific level based on linear measurements from raw and size-adjusted (without the influence of size) data sets. Homo sapiens Pan troglodytes Gorilla gorilla r PLS pE-size SE r PLS pE-size SE r PLS pE-size SE Raw Cr/C1 0.206 0.856 0.999 0.012 0.257 0.488 0.143 0.010 0.552 0.001 4.324 0.012 Cr/C2 0.589 0.009 2.733 0.012 0.774 0.001 5.671 0.010 0.861 0.001 8.562 0.011 Cr/C3 0.537 0.021 2.278 0.011 0.575 0.001 3.609 0.010 0.716 0.001 6.182 0.011 Cr/C4 0.549 0.010 2.739 0.011 0.722 0.001 5.709 0.010 0.862 0.001 8.274 0.011 Cr/C5 0.533 0.018 2.319 0.011 0.707 0.001 5.388 0.010 0.880 0.001 8.792 0.011 Cr/C6 0.534 0.026 2.042 0.012 0.708 0.001 4.975 0.010 0.881 0.001 8.503 0.011 Cr/C7 0.566 0.008 2.745 0.011 0.712 0.001 5.246 0.010 0.866 0.001 8.036 0.011 Size-adjusted Cr/C1 0.364 0.612 0.318 0.014 0.433 0.190 0.894 0.012 0.620 0.004 3.363 0.013 Cr/C2 0.548 0.217 0.768 0.014 0.589 0.068 1.489 0.012 0.664 0.001 3.997 0.012 Cr/C3 0.413 0.678 0.514 0.013 0.511 0.045 1.885 0.010 0.571 0.006 2.787 0.012 Cr/C4 0.450 0.632 0.358 0.013 0.602 0.004 2.905 0.011 0.702 0.001 4.559 0.011 Cr/C5 0.446 0.446 0.102 0.013 0.516 0.090 1.376 0.011 0.704 0.001 4.636 0.011 Cr/C6 0.464 0.306 0.479 0.013 0.407 0.663 0.450 0.011 0.707 0.001 4.613 0.011 Cr/C7 0.434 0.499 0.075 0.013 0.421 0.526 0.106 0.011 0.687 0.001 4.347 0.011 Significant values are indicated in bold (p-value <0.05). Figure 5. Correlation analysis between the cervical vertebrae and the skull from raw data (left) and size-adjusted (right) in Homo sapiens (green), Pan troglodytes (blue), and Gorilla gorilla (light yellow). The numerical values are represented in Table 4. The cervical vertebrae represented in the figure belong to an individual of the species P. troglodytes. Abbreviation: r PLS ¼Partial least square correlation. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 8 results indicate certain differences in the craniocervical correlation pattern between species that present different postural and locomotor repertoires; however, we have not directly tested the relationship between them here. Thus, to confirm our hypothesis that certain differences in postural and locomotor repertoires would be reflected in their patterns of craniocervical integration, further analysis with a more specific approach is needed. Allometry The results showed that all the cervical vertebrae, except the atlas in chimpanzees and humans, exhibited a significant amount of shape variation explained by cranial size (Table 6). Gorillas yielded the largest percentages of variation followed by chimpanzees. Also, the C2eC7 vertebrae presented positive allometric trends; in contrast, the atlas showed a negative trend (Fig. 7). Pairwise analyses yielded similar allometric pattern for the three species in the first (atlas) and last (C7) cervical vertebrae (Table 7). Humans differ from chimpanzees in the C4eC5 vertebrae and from gorillas in the C2eC6. Gorillas and chimpanzees differ in the allometric pattern in the C2 and C5 vertebrae. 3.3. TM analyses of size-adjusted data Compared with the previous section on raw data, the sizeadjusted data give here information on shape without the influence of size. PLS integration analysis The results of the PLS analysis from sizeadjusted databases are presented in Table 4 and Figure 5.In comparison with those results obtained from raw data sets, except for the atlas, all the cervical vertebrae obtained lower covariation values with the cranium. Gorillas displayed significant r PLS values in all the covariation analyses (i.e., cranium with C1eC7) and chimpanzees in C3 and C4 vertebrae, and humans did not show significant results in any craniocervical covariation analyses (Table 4). In consequence, compared with the results obtained from the raw data set, this results in a reduction in the differences in the values of covariation among vertebrae. Also, pairwise comparisons revealed that once the size factor is removed, differences between vertebrae and species are also reduced (SOM Tables S15e16). Finally, these results partially support the hypothesis that removing the effect of size would result in a reduction of the magnitude of integration. Conversely to the rest of the vertebrae, the atlas increased its values. Between-variables pairwise correlation test These results also confirmed this reduction of the levels of correlation from sizeadjusted data sets compared with those obtained from raw data sets, except in the atlas of H. sapiens and P. troglodytes (SOM Table S16 and Fig. S1). Also, in general terms, the highest craniocervical correlation values are not related to variables representing maximum length, either midlateral (e.g., MaxTrDi, STrD, or M8) or dorsoventral (e.g., MaxDVDi, M5, or M13), but variables related to the vertebral foramen (SOM Table S17). 4. Discussion 4.1. Cranium-cervical covariation: differences between species The cervical column of hominoids has been proposed to be highly integrated with the cranial base, with no substantial differences between species (Villamil, 2018). However, our results indicate differences exist both in the patterns of correlation and magnitudes of integration of the cervical vertebrae with the cranium as a whole (Fig. 6). Among hominines, gorillas display the highest magnitude of craniocervical integration in all the vertebrae, followed by chimpanzees. If the magnitude of integration indicates the degree of interaction between the cranium and the cervical vertebrae, the pattern reveals how this is produced (Grabowski et al., 2011). Our results showed that humans differ from both gorillas and chimpanzees in the traits that link the cranium and the cervical spine. In both gorillas and chimpanzees, the strongest correlation occurs between variables/traits representing prognathism and length of the cranial base and maximum length and width of the vertebrae. In contrast, in humans, this is produced Table 5 Highest correlation values between the cranium and cervical variables in Homo sapiens,Pan troglodytes, and Gorilla gorilla from raw data sets. H. sapiens Highest correlation r p2nd Highest correlation r p Cr/C1 Basion-Prosthion/MaxDvDi 0.251 0.113 Zygo-Zygo/STrD 0.237 0.148 Cr/C2 Zygo-Zygo/STrD 0.528 <0.001 Glabella-Inion/M1a 0.518 0.001 Cr/C3 Glabella-Inion/M5 0.513 <0.001 Opisthion-Basion/M11 0.470 0.001 Cr/C4 Glabella-Inion/M8 0.547 <0.001 Nasion-Prosthion/STrD 0.536 <0.001 Cr/C5 Glabella-Inion/M5 0.549 <0.001 Opisthion-Basion/M11 0.525 <0.001 Cr/C6 Nasion-Prosthion/M1 0.546 <0.001 Glabella-Inion/MaxTrDi 0.466 0.002 Cr/C7 Glabella-Inion/STrD 0.538 <0.001 Nasion-Prosthion/MaxTrDi 0.516 <0.001 P. troglodytes Highest correlation r p2nd Highest correlation r p Cr/C1 Opisthion-Basion/M11 0.306 0.051 Opisthion-Basion/MaxTrDi 0.281 0.076 Cr/C2 Basion-Prosthion/MaxTrDi 0.653 <0.001 Basion-Prosthion/MaxDvDi 0.648 <0.001 Cr/C3 Basion-Prosthion/MaxTrDi 0.647 <0.001 Opisthion-Basion/M10 0.516 <0.001 Cr/C4 Basion-Prosthion/MaxDvDi 0.672 <0.001 Nasion-Basion/M5 0.579 <0.001 Cr/C5 Basion-Prosthion/MaxDvDi 0.631 <0.001 Nasion-Basion/StrD 0.600 <0.001 Cr/C6 Basion-Prosthion/MaxDvDi 0.674 <0.001 Nasion-Basion/M5 0.603 <0.001 Cr/C7 Basion-Prosthion/MaxDvDi 0.718 <0.001 Nasion-Basion/M5 0.659 <0.001 G. gorilla Highest correlation r p2nd Highest correlation r p Cr/C1 Glabella-Inion/MaxTrDi 0.621 <0.001 Glabella-Inion/MaxDvDi 0.586 <0.001 Cr/C2 Nasion-Basion/MaxTrDi 0.868 <0.001 Nasion-Basion/MaxDvDi 0.839 <0.001 Cr/C3 Nasion-Basion/MaxTrDi 0.847 <0.001 Basion-Prosthion/MaxTrDi 0.798 <0.001 Cr/C4 Nasion-Basion/MaxTrDi 0.846 <0.001 Basion-Prosthion/MaxDvDi 0.842 <0.001 Cr/C5 Basion-Prosthion/MaxDvDi 0.865 <0.001 Nasion-Basion/MaxDvDi 0.858 <0.001 Cr/C6 Basion-Prosthion/M13 0.862 <0.001 Nasion-Basion/MaxDvDi 0.851 <0.001 Cr/C7 Nasion-Basion/STrD 0.852 <0.001 Basion-Prosthion/MaxDvDi 0.851 <0.001 Significant values are indicated in bold (p-value <0.05). The vertebral dimensions follow Figure 6 and SOM Table S13. M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 9 Lovejoy, C.O., Sewa, G., Spurlock, L., Asfaw, B., White, T.D., 2009b. The pelvis and femur of Ardipithecus ramidus: The emergence of upright walking. Science 326, 71e1e71e6. Lovejoy, C.O., Simpson, S.W., White, T.D., Asfaw, B., Suwa, G., 2009c. Careful climbing in the Miocene: The forelimbs of Ardipithecus ramidus and humans are primitive. Science 326, 70e70e8. Lycett, S.J., Collard, M., 2005. Do homologies impede phylogenetic analyses of the fossil hominids? An assessment based on extant papionin craniodental morphology. J. Hum. Evol. 49, 618e642. Manfreda, E., Mitteroecker, P., Bookstein, F.L., Schaefer, K., 2006. Functional morphology of the first cervical vertebra in humans and nonhuman primates. Anat. Rec. 289B, 184e194. Marroig, G., Cheverud, J.M., 2001. A comparison of phenotypic variation and covariation patterns and the role of phylogeny, ecology, and ontogeny during cranial evolution of New World monkeys. Evolution 55, 2576e2600. Marroig, G., Shirai, L.T., Porto, A., de Oliveira, F.B., De Conto, V., 2009. The evolution of modularity in the mammalian skull II: Evolutionary consequences. Evol. Biol. 36, 136e148. Martin, R., Saller, K., 1957. Lehrbuch der Anthropologie. Gustav Fischer, Stuttgart. McCown, T.D., Keith, A., 1939. The Stone Age of Mount Carmel. The Fossil Human Remains from the Levalloiso-Mousterian. Clarendon Press, Oxford. Melo, D., Garcia, G., Hubbe, A., Assis, A.P., Marroig, G., 2016. EvolQR: An R package for evolutionary quantitative genetics. F1000Research 4, 1e25. Meyer, M.R., Woodward, C., Tims, A., Bastir, M., 2018. Neck function in early hominins and suspensory primates: Insights from the uncinate process. Am. J. Phys. Anthropol. 166, 613e637. Mitteroecker, P., Bookstein, F., 2007. The conceptual and statistical relationship between modularity and morphological integration. Syst. Biol. 56, 818e836. Müller, M.A., Merten, L.J., B€ ohmer, C., Nyakatura, J.A., 2021. Pushing the boundary? Testing the ‘functional elongation hypothesis’of the giraffe’s neck. Evolution 75, 641e655. Nalley, T.K., Grider-Potter, N., 2015. Functional morphology of the primate head and neck. Am. J. Phys. Anthropol. 156, 531e542. Nalley, T.K., Grider-Potter, N., 2017. Functional analyses of the primate upper cervical vertebral column. J. Hum. Evol. 107, 19e35. Nalley, T.K., Grider-Potter, N., 2019. Vertebral morphology in relation to head posture and locomotion I: The cervical spine. In: Been, E., G omez-Olivencia, A., Kramer, P.A. (Eds.), Spinal Evolution: Morphology, Function, and Pathology of the Spine in Hominoid Evolution. Springer International Publishing, Cham, pp. 35e50. Narita, Y., Kuratani, S., 2005. Evolution of the vertebral formulae in mammals: A perspective on developmental constraints. J. Exp. Zool. Part B Mol. Dev. Evol. 304B, 91e106. Olson, E.C., Miller, R.L., 1958. Morphological Integration. The University of Chicago Press, Chicago. Pablos, A., Martínez, I., Lorenzo, C., Gracia, A., Sala, N., Arsuaga, J.L., 2013. Human talus bones from the Middle Pleistocene site of Sima de los Huesos (Sierra de Atapuerca, Burgos, Spain). J. Hum. Evol. 65, 79e92. Pontzer, H., Holloway, J.H., Raichlen, D.A., Lieberman, D.E., 2009. Control and function of arm swing in human walking and running. J. Exp. Biol. 212, 523e534. Porto, A., de Oliveira, F.B., Shirai, L.T., De Conto, V., Marroig, G., 2009. The evolution of modularity in the mammalian skull I: Morphological integration patterns and magnitudes. Evol. Biol. 36, 118e135. Porto, A., Shirai, L.T., de Oliveira, F.B., Marroig, G., 2013. Size variation, growth strategies, and the evolution of modularity in the mammalian skull. Evolution 67, 3305e3332. Powell, V., Esteve-Altava, B., Molnar, J., Villmoare, B., Pettit, A., Diogo, R., 2018. Primate modularity and evolution: First anatomical network analysis of primate head and neck musculoskeletal system. Sci. Rep. 8, 1e10. Preuschoft, H., 2004. Mechanisms for the acquisition of habitual bipedality: Are there biomechanical reasons for the acquisition of upright bipedal posture? J. Anat. 204, 363e384. Preuschoft, H., Klein, N., 2013. Torsion and bending in the neck and tail of sauropod dinosaurs and the function of cervical ribs: Insights from functional morphology and biomechanics. PLoS One 8, e78574. Profico, A., Piras, P., Buzi, C., Di Vincenzo, F., Lattarini, F., Melchionna, M., Veneziano, A., Pasquale, R., Manzi, G., 2017. The evolution of cranial base and face in Cercopithecoidea and Hominoidea: Modularity and morphological integration. Am. J. Primatol. 79, e22721. R Core Team, 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL. https://www.Rproject.org/. Randau, M., Cuff, A.R., Hutchinson, J.R., Pierce, S.E., Goswami, A., 2017. Regional differentiation of felid vertebral column evolution: A study of 3D shape trajectories. Org. Divers. Evol. 17, 305e319. Robinson, J.T., 1972. Early Hominid Posture and Locomotion. The University of Chicago Press, Chicago. Rohlf, F.J., Slice, D., 1990. Extensions of the Procrustes method for the optimal superimposition of landmarks. Syst. Biol. 39, 40e59. Rohlf, F.J., Corti, M., 2000. Use of two-block partial least-squares to study covariation in shape. Syst. Biol. 49, 740e753. Ryan, T.M., Sukhdeo, S., 2016. KSD-VP-1/1: Analysis of the postcranial skeleton using high-resolution computed tomography. In: Haile-Selassie, Y., Su, D.F. (Eds.), The Postcranial Anatomy of Australopithecus afarensis. Springer, Dordrecht, pp. 39e62. Ryan, T.M., Carlson, K.J., Gordon, A.D., Jablonski, N., Shaw, C.N., Stock, J.T., 2018. Human-like hip joint loading in Australopithecus africanus and Paranthropus robustus. J. Hum. Evol. 121, 12e24. Schluter, D., 1996. Adaptive radiation along genetic lines of least resistance. Evolution 50, 1766e1774. Schultz, A.H., 1942. Conditions for balancing the head in primates. Am. J. Phys. Anthropol. 29, 483e497. Schultz, A.H., 1961. Primatologia. Handbuch der primatenkunde. S. Karger, Basel. Sidlauskas, B., 2008. Continuous and arrested morphological diversification in sister clades of characiform fishes: A phylomorphospace approach. Evolution 62, 3135e3156. Slijper, E., 1946. Comparative biologic anatomical investigations on the vertebral column and spinal musculature of mammals. Tweede Sect. 17, 1e128. Sockol, M.D., Raichlen, D.A., Pontzer, H., 2007. Chimpanzee locomotor energetics and the origin of human bipedalism. Proc. Natl. Acad. Sci. USA 104, 12265e12269. Steppan, S.J., Phillips, P.C., Houle, D., 2002. Comparative quantitative genetics: Evolution of the G matrix. Trends Ecol. Evol. 17, 320e327. Stern Jr., J.T., 2000. Climbing to the top: A personal memoir of Australopithecus afarensis. Evol. Anthropol. 9, 113e133. Strait, D.S., 2001. Integration, phylogeny, and the hominid cranial base. Am. J. Phys. Anthropol. 114, 273e297. Strait, D.S., Ross, C.F., 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, 205e222. Toerien, M.J., 1957. Note on the cervical vertebrae of the La Chapelle man. S. Afr. J. Sci. 53, 447e449. Toerien, M.J., 1961. The length and inclination of the primate cervical spinous processes. Trans. R. Soc. South Africa 36, 95e105. Vander Linden, A., Campbell, K.M., Bryar, E.K., Santana, S.E., 2019a. Head-turning morphologies: Evolution of shape diversity in the mammalian atlaseaxis complex. Evolution 73, 2060e2071. Vander Linden, A., Hedrick, B.P., Kamilar, J.M., Dumont, E.R., 2019b. Atlas morphology, scaling and locomotor behaviour in primates, rodents and relatives (Mammalia: Euarchontoglires). Zool. J. Linnean Soc. 185, 283e299. Varela-Lasheras, I., Bakker, A.J., van der Mije, S.D., Metz, J.A., van Alphen, J., Galis, F., 2011. Breaking evolutionary and pleiotropic constraints in mammals: On sloths, manatees and homeotic mutations. EvoDevo 2, 11. Var on-Gonz alez, C., Whelan, S., Klingenberg, C.P., 2020. Estimating phylogenies from shape and similar multidimensional data: Why it is not reliable. Syst. Biol. 69, 863e883. Villamil, C.I., 2018. Phenotypic integration of the cervical vertebrae in the Hominoidea (Primates). Evolution 72, 490e517. Villamil, C.I., 2021. The role of developmental rate, body size, and positional behavior in the evolution of covariation and evolvability in the cranium of strepsirrhines and catarrhines. J. Hum. Evol. 151, 102941. Vidal, P.P., Graf, W., Berthoz, A., 1986. The orientation of the cervical vertebral column in unrestrained awake animals. Exp. Brain Res. 61, 549e559. von Cramon-Taubadel, N., Smith, H.F., 2012. The relative congruence of cranial and genetic estimates of hominoid taxon relationships: Implications for the reconstruction of hominin phylogeny. J. Hum. Evol. 62, 640e653. Wagner, G.P., 1988. The influence of variation and of developmental constraints on the rate of multivariate phenotypic evolution. J. Evol. Biol. 1, 45e66. Wagner, G.P., 1996. Homologues, natural kinds and the evolution of modularity. Am. Zool. 36, 36e43. Wagner, G.P., Schwenk, K., 2000. Evolutionarily stable configurations: Functional integration and the evolution of phenotypic stability. Evol. Biol. 31, 155e217. Warrener, A.G., Lewton, K.L., Pontzer, H., Lieberman, D.E., 2015. A wider pelvis does not increase locomotor cost in humans, with implications for the evolution of childbirth. PLoS One 10, e0118903. Watanabe, A., Fabre, A.C., Felice, R.N., Maisano, J.A., Müller, J., Herrel, A., Goswami, A., 2019. Ecomorphological diversification in squamates from conserved pattern of cranial integration. Proc. Natl. Acad. Sci. USA 116, 14688e14697. Wellik, D.M., Capecchi, M.R., 2003. Hox10 and Hox11 genes are required to globally pattern the mammalian skeleton. Science 301, 363e367. White, A.A., Panjabi, M.M., 1990. Clinical Biomechanics of the Spine. J. B. Lippincott, Philadelphia. Whitcome, K.K., Shapiro, L.J., Lieberman, D.E., 2007. Fetal load and the evolution of lumbar lordosis in bipedal hominins. Nature 450, 1075e1078. Zelditch, M.L., 1988. Ontogenetic variation in patterns of phenotypic integration in the laboratory rat. Evolution 42, 28e41. Zubair, H.N., Chu, K.M., Johnson, J.L., Rivers, T.J., Beloozerova, I.N., 2019. Gaze coordination with strides during walking in the cat. J. Physiol. 597, 5195e5229. M. Arlegi, A. Pantoja-P erez, C. Veschambre-Couture et al. Journal of Human Evolution 162 (2022) 103112 16