• human evolution;
  • size and shape;
  • complex morphology;
  • retromolar space;
  • chin


  1. Top of page
  2. Abstract
  7. Acknowledgements

Allometry is an important factor of morphological integration that contributes to the organization of the phenotype and its variation. Variation in the allometric shape of the mandible is particularly important in hominid evolution because the mandible carries important taxonomic traits. Some of these traits are known to covary with size, particularly the retromolar space, symphyseal curvature, and position of the mental foramen. The mandible is a well studied system in the context of the evolutionary development of complex morphological structures because it is composed of different developmental units that are integrated within a single bone. In the present study, we investigated the allometric variation of two important developmental units that are separated by the inferior nerve (a branch of CN V3). We tested the null hypothesis that there would be no difference in allometric variation between the two components. Procrustes-based geometric morphometrics of 20 two-dimensional (2D) landmarks were analyzed by multivariate regressions of shape on size in samples from 121 humans, 48 chimpanzees, and 50 gorillas (all recent specimens), eight fossil hominids from Atapuerca, Sima de los Huesos (AT-SH), and 17 Neandertals. The findings show that in all of the examined species, there was significantly greater allometric variation in the supra-nerve unit than in the infra-nerve unit. The formation of the retromolar space exhibited an allometric relationship with the supra-nerve unit in all of the species studied. The formation of the chin-like morphology is an “apodynamic” feature of the infra-nerve unit in the AT-SH hominids. The results of this study support the hypothesis that allometry contributes to the organization of variation in complex morphological structures. Anat Rec Part A 278A:551–560, 2004. © 2004 Wiley-Liss, Inc.

In evolutionary studies of genetics, development, and comparative morphology, the mammalian mandible is the paradigmatic representative of complex morphological structures (Atchley and Hall, 1991; Klingenberg and Leamy, 2001; Klingenberg, 2002). Such structures are characterized by a series of different and interacting units and processes that, in sum, are responsible for the final morphology of the object. In addition, a detailed understanding of mandibular morphology is crucial for interpreting human evolution, both because the mandible is one of the best-preserved bones in the hominid fossil record, and because it carries a series of systematically relevant traits (Rosas, 1997, 2001).

The mandibles of the Atapuerca, Sima de los Huesos (AT-SH), hominids (Fig. 1) represent an exceptional fossil sample from the Middle Pleistocene hominid record (Rosas, 1995, 1997). They are interesting not only because of their remarkable state of conservation, but also because they provide researchers a unique opportunity to study morphological variation within a single Middle Pleistocene hominid population (Bermúdez de Castro et al., 2001). Recent radiometric dating has suggested a possible age for these hominids of 350–500 kyr (Bischoff et al., 2002).

thumbnail image

Figure 1. Two mandibles representing the variability from the AT-SH Middle Pleistocene site. Left: The large specimen (AT-605) is probably a male. Right: The small specimen (AT-505, AT-604, and AT-952) is considered to be a female.

Download figure to PowerPoint

From a biological perspective, a number of studies have suggested that the morphological variation in the AT-SH hominids follows essentially an allometric trajectory (Rosas, 1992, 1997, 2000). A major component of shape variation in these mandibles is thus related to size, or the attainment of adult size (i.e., growth). Various possible origins of the strong allometric nature of this morphological variation have been proposed, such as particular patterns of growth and development, and specific patterns of sexual dimorphism (Rosas, 2000; Rosas et al., 2002a); however, the evolutionary nature and consequences of this allometry are still not fully understood.

It is important to gain a deeper understanding of this allometric variation not only to increase our knowledge regarding growth and development in craniofacial biology and paleobiology, but with respect to evolutionary biology as well. This is because some morphological traits (e.g., the presence of a retromolar space, the curvature and orientation of the symphysis, and the position of the mental foramen with respect to the dentition) that show a clearly positively allometric association with size in the AT-SH hominids are also characteristic apomorphic features that foreshadow later classic Neandertal traits. Thus, they are of specific systematic value in hominid paleontology (Rosas, 1992; Franciscus and Trinkaus, 1995; Rosas, 2001).

From a methodological viewpoint, Procrustes-based geometric morphometry of landmark data is an ideal tool for the analysis of allometry, for two reasons (Bookstein, 1991): First, landmarks can be used as the carrier of biological hypotheses of different morphogenetic units, and their patterns of variation can be analyzed in isolated as well as pooled data sets. Second, these methods are based on the operational separation of size and shape, which provides a surprisingly high analytical resolution in questions of allometry (Bookstein, 1991, 1996; Rosas and Bastir, 2002).

The aim of the present study was to perform a detailed comparative morphological analysis of allometric patterns of shape variation in mandibles from modern humans, fossil hominids, and extant hominoids by geometric morphometry. Specifically, we explored allometric variation patterns and their possible evolutionary modifications in the dissociation or integration of mandibular developmental units that may have been involved in hominid evolution in Middle Pleistocene Europe.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Developmental Units and Morphological Evolution

The mandible is a paradigmatic key object for studies of complex morphology, in which morphological structure is characterized as being “comprised by a number of developmental parts and processes” (Atchley and Hall, 1991). The final adult morphology is considered to be the result of a morphogenetic choreography (Atchley and Hall, 1991) that integrates the dynamics of a variety of biological processes and structures, their principal effects, and their interactions and mutual limitations (Maynard Smith et al., 1985; Alberch, 1990).

There are different hierarchical levels in the organization of a phenotype. At the embryological level of developmental hierarchy, the mandible is composed of five different morphogenetic components: the alveolar part, Meckel's cartilage (around which will be formed the core of the mandible), and the angular, condylar, and coronoid processes (Atchley and Hall, 1991). Another level of morphological organization is the classical anatomic separation into the ramus and corpus. Recent analyses have supported the hypothesis that the ramus and the corpus are distinct developmental modules (Enlow and Hans, 1996; Klingenberg et al., 2003). A distinct pattern of morphological covariation becomes particularly evident in the analysis of shape variation that is corrected for allometry and fluctuating asymmetry (Klingenberg et al., 2003).

At the next higher level of integration, which includes allometric variation (Chernoff and Magwene, 1999) as the final morphological effect of postnatal growth and development, the mandible can be divided into different units (Fig. 2). At this level, the structures superior to the inferior alveolar nerve canal differ from osteological structures inferior to it, with respect to bone remodeling (i.e., bone resorption and bone deposition) (Enlow and Hans, 1996). It is known that in humans the unit superior to the alveolar nerve canal is characterized ontogenetically by external resorption and internal deposition, while at the infra-alveolar component the external surface is characterized by deposition, and the internal surface is resorptive (Enlow and Hans, 1996).

thumbnail image

Figure 2. Generalized hominid mandible in which one can distinguish the supra-nerve (shaded area) and infra-nerve units of the bone. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

Finally, to a certain extent, this “infra-nerve” unit contains reflections of the growth process and the properties of mandibular growth rotations (e.g., a complete mandibular, and matrix and intramatrix rotations (sensu Björk, 1969, 1991)). Björk (1969) proposed that the structures of the infra-alveolar system (i.e., the mandibular flexure, the anteroposterior position of the inferior basal border at the corpus, and the associated vertical position), and the orientation and configuration of the anterior symphysis are direct morphological reflections of growth rotations and the locations of the rotation centers. Morphologically, these structures are remarkably constant in postnatal ontogeny. All of the morphological structures of the infra-nerve unit can be inspected at the outer mandibular profile.

The “supra-nerve” mandibular unit (Enlow and Hans, 1996), which is superior to the alveolar nerve, is developmentally different from the infra-nerve unit. This region is also strongly related to occlusion and the masticatory system, and consists of the insertion site of the temporalis muscle, the dentition, and the alveolar process. Therefore, most structures in this system are located along the anterior border of the ramus (e.g., the microskeletal unit of the coronoid process and temporal muscles (Moss and Young, 1960; Enlow and Hans, 1996)) and the superior border of the corpus, which is, in sum, the upper mandibular profile. Since they are important parts of the viscerocranium, one can expect their size and shape to relate strongly to variation in body size (Enlow and Hans, 1996; Rosas and Bastir, 2002).

Study Design

The composite nature of these morphological structures can be reflected by a corresponding analytical design. Thus, by performing a separate allometric analysis of these different developmental and functional units, it becomes theoretically possible to focus questions regarding the particular allometry of the AT-SH mandibles directly on evolutionary modifications of patterns of growth and development, and to observe the evolutionary developmental trajectory of the analyzed regions. Further, the groups analyzed must be closely related evolutionarily. Thus, it is most logical to compare the fossil hominid sample to the African great apes, the closest living relatives of modern humans.

In the present study, we test the null hypothesis that there is no difference in between the supra- and infra-nerve developmental units in terms of allometric variation. In the first step, the allometric patterns of the complete mandibles from five different species were analyzed quantitatively and visually from a confirmatory and exploratory perspective. We investigated how species differences in allometry are expressed morphologically in a multivariate, geometric size-shape manner. In a further step, we analyzed different structural and functional allometry patterns in subsets of mandibular developmental regions in a comparative interspecific framework by multivariate regression analyses (Rohlf, 1998a; Penin et al., 2002).


  1. Top of page
  2. Abstract
  7. Acknowledgements

The human samples investigated in this work originated from two geographically different populations, and are described in detail elsewhere (Rosas and Bastir, 2002; Bastir et al., 2004; Bastir et al., in press). The European sample (N = 82) was housed at the Institute of Anthropology at the University of Coimbra in Portugal. The African sample (N = 39) was housed at the Natural History Museum (NHM) in London, as were the chimpanzee (N = 48) and gorilla (N = 50) samples. The original mandibles from the AT-SH (N = 8) were measured at the Museo Nacional de Ciencias Naturales (MNCN), Madrid. Some of the Neandertal mandibles (N = 17) were measured on casts at the MNCN, and on casts and one original (Tabun 2) at the NHM. All of the extant samples included a comparable number of males and females. The fossils are listed in Table 1.

Table 1. Fossil specimens included in the analyses*
Atapuerca SH sampleNeandertal sample
  • *

    Sex attibution after Rosas et al. (2002a). The new specimen AT-3888 was considered female.

AT-300MaleRight, mandibular bodyLa ChapelleAmud 1
AT-605MaleRight ramus missingAubesierTabun 1
AT-888MaleLeft condylar part missingMonte Circeo 3Tabun 2
AT-607FemaleLeft ramus missingLa Ferassie1Bañolas
AT-950FemaleComplete mandibleLa Quina 5Zafarraya
AT-952FemaleComplete mandibleSt. CesaireMonte Circeo 2
AT-2193FemaleRight hemimandibleLa Quina 9Spy 1
AT-3888FemaleRight hemimandibleKebara 

A large, homogeneous sample of hominid remains was recovered from the Sima de los Huesos site in the Atapuerca Hills of northern Spain. All of the human bones were deposited during the same sedimentation period (Bischoff et al., 1997) (see Arsuaga et al. (1997) for details regarding the SH site). The AT-SH Middle Pleistocene sample is exceptional: not only does it contain a large number of specimens (>4,000), it also represents a minimum of 28 individuals of the same biological population (Bermúdez de Castro et al., 2004). Recent radiometric studies (U-series) of a 14-cm-thick in situ speleotheme overlying the mud-breccia containing the human bones determined a minimum age of 350,000 years for the SH hominins (Bischoff et al., 2002). Estimations of the speleotheme growth rate, together with other contextual evidence, indicate an interval of 400,000–500,000 years (oxygen isotopic stages 12–14) for these hominins (Bishoff et al., 2002). These hominins have been included in the species H. heidelbergensis (Arsuaga et al., 1997). Regarding the specific mandibular data, although the total sample contained at least 29 different specimens (Rosas, 1995, 1997, 2000), eight of these mandibles were almost complete and thus were selected for this study (see Table 1).

A series of 3D landmarks were digitized by means of a Microcribe 3DX digitizer (Immersion Corporation, San Jose, CA) (Table 2). The three-dimensional (3D) data were then transformed a two-dimensional (2D) coordinated set, which we analyzed by common geometric morphometry (Bookstein, 1991). The data conversion from 3D into 2D was necessary for the thin-plate splines analyses (Bookstein, 1991). Bilateral landmarks were averaged, and missing landmarks were replaced by the group- and/or sex-specific mean coordinates (Rosas and Bastir, 2002). The sex attribution in the Neandertal sample was not clear, and thus the missing landmarks were replaced by the grand mean values of the corresponding point.

Table 2. List of the digitized 3D landmarks*
  • *

    Rosas and Bastir (2002) except where additionally defined.

M3Mesial to M3, alveolar border
M1Mesial to M1, alveolar border
CanineMesial to C, alveolar border
Mental foramen 
Inferior basal border 
Preangular notch 
Ramus flexure 
Mandibular notchMaximum flexion of curvature
CoronoidAnterior-superior tip of coronoid process
Anterior ramusAnterior point of minimal ramus breadth
Retromolar spacePosterior limit of alveolar process
Mandibular foramen 
Internal infradentale 

Geometric Morphometry

Procrustes-based geometric morphometry is based on the analysis of landmark coordinates. At the core of this method is the separation between two components of form (i.e., size and shape) (Bookstein, 1991). Shape is the residual geometric information that remains when differences due to location, scale, and rotational effects are removed (Kendall, 1977). Procrustes superimposition techniques minimize the offset between homologue landmarks, and size is obtained as a scaling factor (termed the “centroid size”) (Rohlf and Slice, 1990; Bookstein, 1991).

The thin-plate spline decomposition method (Bookstein, 1991; Rohlf, 1996; Rohlf et al., 1996) is used to produce partial warp and uniform component scores for further analysis. These scores represent all information about the shape of the specimens with the same number of variables as degrees of freedom in the shape measurement. In the present study, the Procrustes mean shape of all of the specimens is used as the reference form for the thin-plate spline decomposition. The overall variation in shape is small enough that common statistical procedures may be used to analyze shape data in the Euclidean tangent space to Kendall's shape space (Bookstein, 1991, 1996; Rohlf, 1996; Rohlf et al., 1996; Dryden and Mardia, 1998; Slice, 2001). This was tested with the use of tpsSMALL (Rohlf, 1998a), and the correlation between shape distances in Kendall's shape space and tangent space was 0.9999.

Relative Warps

Relative warps are the principal components of shape variables (such as Procrustes residuals or partial warps and uniform component scores), and reflect the major patterns of shape variation within a group (Bookstein, 1991; Rohlf, 1996; Rohlf et al., 1996). They were analyzed with the use of tpsRELW (α = 0) (Rohlf, 1998c).

Quantitative Analyses

First, we performed a multivariate analysis of covariance (MANCOVA) on the pooled sample to test for similarity of allometric patterns (Bookstein, 1996) using StatSoft (STATISTICA for Windows; StatSoft, Inc., Tulsa, OK), and then we conducted a relative warp analysis (Rohlf, 1998b). Second, for the complete mandible, the supra- and the infra-nerve units, and each species, we performed separate regression analyses using tpsREGR (Rohlf, 1998a). The fit of the each of the five species-specific models, as well as their statistical significance, were evaluated by the percentage of explained variance and the associated statistics given by the software (Rohlf, 1998a).

We evaluated the similarity of regression slopes using VecCompare (IMP software) (Sheets, 2001; Zelditch et al., 2003). This program compares the between-group angle with the upper 95% C.I. of within-group angle ranges assessed by a bootstrapping approach (N = 300). It tests the null hypothesis that a given range of angles could have arisen by random sampling of one group. If the between-group angle is larger than the within-group angles, that difference is significant at the α = 0.05 level. The bootstrapping approach is performed to account for differences in sample size between populations. Three possible results can be obtained: If the between-group angle is larger than both within-group ranges, the slopes are significantly different. If the between-group angle is larger than one within-group range, the bootstrapping approach is not informative, because while the one sample yields smaller ranges than the between-group range, the other group with the large range could still cover the angle range between groups. If the between-group range is smaller than both within-group ranges, no difference can be inferred in the regression slopes.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Multivariate Patterns of Allometry in Complete Mandibles

The overall variation in the mandibular morphology of all five species examined is shown in Figure 3. Table 3 shows that the allometric patterns significantly differ among the species, as indicated by the statistically significant interaction term of the MANCOVA model. Consequently, we analyzed the allometric shape variation by separate Procrustes and regression analyses in each species, using tpsREGR (Rohlf, 1998b) (Table 4).

thumbnail image

Figure 3. The first two principal components of shape data (i.e., the relative warps) reflect the major patterns of morphological variation in the pooled sample. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

Table 3. MANCOVA, partial warps scores and uniform components of complete mandibular shape data*
  • *

    16 landmarks, factor: five species'; covariate: centroid size; interaction term: allometry within species.

Centroid sizeWilks0.6313.49628168.000.000
Species centroid sizeWilks0.4501.330112669.810.018
Table 4. Multivariate regressions of shape on size for each component and each species*
 Regression fit (% expl. variance)P-levels (Goodall's F or Wilks λ)
  • *

    Percent of explained variance and significance levels (P) are given.


The visualizations depicted in Figure 4 show the smallest and largest specimens of each species. In all of the mandibles it can be seen that the larger specimens show significant variation at the alveolar region associated with the retromolar area. This effect is smallest (although present) in humans (Fig. 4a), and is well represented in the AT-SH and Neandertal mandibles (Fig. 4d and e). Interestingly, the allometric effect on the retromolar area is also present in the African apes (Fig. 4b and c). Table 4 shows that the total mandibular allometry was statistically significant in all of the species except the chimpanzees and Neandertals.

thumbnail image

Figure 4. Individual regressions of shape on centroid size visualize the species-specific patterns of allometric shape variation within the observed size range. The left and right columns shows the smallest and largest individuals, respectively, of the sample. All of the species show significant variation at the retromolar area, between the ramus and corpus and their relative proportions. a: Humans. b: Chimpanzees. c: Gorillas. d: AT-SH hominids. e: Neandertals.

Download figure to PowerPoint

Allometries of Developmental Mandibular Units

The allometric patterns of shape variation of the supra-nerve unit are depicted in Figure 5. The morphological allometric variation shows the corresponding morphological effects observed in the allometries of the complete mandibles (Fig. 4). Table 4 shows that all of the regressions were highly significant and accounted for more variance than the total mandibular models. The allometric effects on morphological variation were much stronger in the supra-nerve unit than in the infra-nerve unit (Fig. 6).

thumbnail image

Figure 5. Allometries of the developmental units. Observed size range; smallest and largest individuals of each species. Note the modifications in the relative position of the ramus and corpus, the modifications of the retromolar area at the supra-nerve unit, and the symphyseal modification in the infra-nerve unit of hominids and humans. a: Humans. b: Chimpanzees. c: Gorillas. d: AT-SH hominids. e: Neandertals.

Download figure to PowerPoint

thumbnail image

Figure 6. Explained variance of regressions of shape on size for the supra- and infra-nerve units. Note the significantly higher allometric variation in the supra-nerve unit. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

We further analyzed the significant supra-nerve units by bootstrapping (N = 300) the angles of the regressions. The results did not yield clear indications regarding differences in the slopes of the regression models. In those comparisons in which one intragroup range was lower than the between-group range, the other intragroup range was larger which indicated that the results were not informative regarding the tested hypotheses (Table 5). Therefore, the analyses were based on the visual interpretations of TPS-regression transformation grids.

Table 5. Angles and angle ranges in degrees*
  • *

    Bootrapping analyses; N = 300; VecCompare, IMP-software Sheets (2001) for the significant regressions of the supra-alveolar unit.


In general, the regression models of the infra-nerve units accounted for very low fractions of variance, and were significant only for humans and gorillas. Morphologically, however, they support the hypothesis that apomorphic traits exist in large AT-SH hominids (Figs. 4d and 5d).


  1. Top of page
  2. Abstract
  7. Acknowledgements

In the present study we investigated allometric mandibular variation in AT-SH hominids in a broad interspecific comparative framework of modern humans and African great apes. The mandible was divided into two developmental units separated by the infra-alveolar nerve. We tested the null hypothesis that there is no difference in allometric variation between the supra- and infra-nerve units. The present findings indicate that the null hypothesis should be rejected. The morphological variation of the supra-nerve unit showed a significantly stronger allometric component than the infra-nerve unit.

Biological Scope of Allometry

Allometric analyses are important for a number of reasons. There are three different kinds of size-shape relationships that should be distinguished (i.e., ontogenetic, static, and evolutionary allometry) (Cheverud, 1982). All of these are interrelated because ontogenetic allometries produce the variation that is present in adult allometric variation and becomes modified during evolution (Klingenberg, 1998).

The correspondence between adult and ontogenetic allometry is not direct, but is often “at least fairly close” (Klingenberg, 1998). It appears that static allometry in humans reflects variation in skull proportions, i.e., positive allometry of the face and negative allometry of the braincase, which is similar to the effects of later postnatal growth (Godfrey and Sutherland, 1996; Rosas and Bastir, 2002). Considering that static allometric variation reflects ontogenetic relations, it may be equally informative in evolutionary comparisons. Similar ontogenetic allometries have been suggested to indicate phylogenetic proximity (Ponce de León and Zollikofer, 2001; Ackermann and Krovitz, 2002).

In morphological terms, allometry is a large-scale integrative factor (Chernoff and Magwene, 1999). Allometry contributes to the organization of the phenotype, because it establishes the adult proportions as it integrates body size with the relative size of the viscerocranium. This allometric correspondence has been associated functionally with the masticatory and respiratory system, since both of these systems relate biologically to the size of the body (Emerson and Bramble, 1993; Smith, 1993; Enlow and Hans, 1996).

However, allometric variation may also contribute to our understanding of patterns of morphological variation in the mandible. This is particularly evident in the results of this study. Our findings show that in all five species examined, the supra-nerve unit displayed strong allometric variation that accounted for higher variance than the allometry of the infra nerve unit (Fig. 6).

This is illustrated in Figure 4, which shows a strong allometric variation in the posterior border of M3. It is possible to interpret these variation patterns as being morphologically expressed on a given bauplan of the skull, which is either more horizontal or more vertical. The allometric formation of the retromolar space appears to follow this architectonic design, and determines to some degree whether the trait (which is potentially present in all larger individuals) becomes recognizable.

Morphological allometric variation in humans is minimal, and is known to be expressed mainly in the vertical direction (Rosas and Bastir, 2002). This variation is reflected in the modifications of the increased vertical ramus proportions compared to those of the horizontal corpus in larger humans (Fig. 4a). In contrast, the variation in chimpanzees is organized in a more horizontal skull design (Fig. 4b). The retromolar space formation reflects this principal skull orientation. However, gorillas express retromolar variation in a strong vertical direction, given the extreme proportions of the viscerocranium, which is represented in the relative proportions of the ramus and corpus in the allometric variation pattern (Fig. 4c).

With respect to the hominid fossils, the formation of the retromolar space is of specific relevance because this trait has been used as a diagnostic indicator of Neandertal evolution (Franciscus and Trinkaus, 1995; Rosas, 1997, 2001). The comparative evidence from the African great apes, as discussed above, offers researchers an opportunity to analyze the formation of the hominid retromolar space in terms of skull design.

It is well known that a key trend in human evolution is the reduction of prognathism and/or facial projection (Aiello and Dean, 1990). Whatever the reason for these evolutionary tendencies (Lieberman et al., 2002), the present results suggest that there is a tendency toward increasing verticality, as indicated by the horizontal and higher allometric variation in AT-SH hominids (Fig. 4d) and Neandertals (Fig. 4e) compared to the vertical and lower variation among modern humans (Fig. 4a). These variation patterns at the retromolar area appear to follow the trends observed in the evolution of the relative proportions of the braincase and the face (Lieberman et al., 2002).

Another interesting tendency is the allometric modification of the anterior symphyseal curvature, which may be considered together with the verticality or horizontality of the hominid face. In the AT-SH hominids, the morphological array ranges from being absent in smaller individuals to being clearly pronounced in larger individuals. In Neandertals, the curvature is less variable but is always slightly present, and it varies particularly when the height of the part inferior to the symphyseal depression is increased in large individuals. In humans, the curvature is always well pronounced, and in large individuals the symphysis changes its orientation into a slightly more upright position. Is the curvature in the hominids thus homologous to the chin of humans, as the allometric variation patterns appear to indicate? It is difficult to understand this morphological pattern, because although remodeling patterns in humans show a constant distribution of deposition and resorption (Enlow, 1968), recent analyses of AT-SH hominids indicate positional variation in the local distribution of these osteogenetic fields (Martinez-Maza and Rosas, 2002).

It has been hypothesized that the human chin is as a consequence of vertical, orthognathic skull design and particular growth-related mandibular rotations (Enlow and Hans, 1996). Middle Pleistocene hominids do not show the human degree of orthognathism, so this explanation may not be entirely valid for the fossil mandibles. Further studies, particularly from an osteogenetic perspective, are necessary to address this difficult problem.

A general conclusion of the present study is that a morphological character appears to be the result of a specific biological process, which is expressed upon a specific bauplan. While the same process in a different morphological context may produce different results (e.g., allometry of the supra-nerve unit), it is possible that a similar morphology could be produced by different processes (symphyseal curvature).


  1. Top of page
  2. Abstract
  7. Acknowledgements

We thank Eugenia Cunha (University of Coimbra), Louise Humphrey, Robert Kruszynski, Paula Jenkins, and Chris Stringer (all NHM-London) for permission to study the materials in their charge. This study was partly supported by SYS Resources of NHM-London, and the Luso-Español Project, Acción Integrada (grant HP1998-0031 to M.B.). The materials and the subject of this investigation were included in the framework of projects BOS2003-08938-CO3-02 and BOS2003-01531 of the Dirección General de Investigación of the Spanish government. M.B. is funded by an FPI predoctoral fellowship from the Spanish government (MCYT).


  1. Top of page
  2. Abstract
  7. Acknowledgements
  • Ackermann RR, Krovitz GE. 2002. Common patterns of facial ontogeny in the hominid lineage. Anat Rec 269: 142147.
  • Aiello L, Dean C. 1990. An introduction to human evolutionary anatomy. London: Academic Press/Harcourt Brace & Company. 596 p
  • Alberch P. 1990. Natural selection and developmental constraints: external versus internal determinants of order in nature. In: DeRousseauCJ, editor. Primate life history and evolution. New York: Wiley-Liss. p 1535.
  • Arsuaga JL, Gracia A, Lorenzo C. 1997. The Sima de los Huesos crania (Sierra de Atapuerca, Spain). A comparative study. J Hum Evol 33: 219281.
  • Atchley WR, Hall BK. 1991. A model for development and evolution of complex morphological structures. Biol Rev 66: 101157.
  • Bastir M, Rosas A, Kuroe K. 2004. Petrosal orientation and mandibular ramus breadth: evidence of a developmental integrated petroso-mandibular unit. Am J Phys Anthropol 123: 340350.
  • Bastir M, Rosas A, Sheets DH. The morphological integration of the hominoid skull: a partial least squares and PC analysis with morphogenetic implications for European Mid-Pleistocene mandibles. In: SliceD, editor. Developments in primatology: progress and prospects. Kluwer Academic/Plenum (in press).
  • Bermúdez de Castro JM, Sarmiento S, Cunha E, Rosas A, Bastir M. 2001. Dental size variation in the Atapuerca-SH middle pleistocene hominids. J Hum Evol 41: 195209.
  • Bermúdez de Castro JM, Martinon-Torres M, Sarmiento S, Lozano M, Arsuaga J-L, Carbonell E. 2004. Rates of anterior tooth wear in middle pleistocene hominins from Sima de los Huesos (Sierra de Atapuerca, Spain). Proc Natl Acad Sci U S A 100: 1199211996.
  • Bischoff JL, Fitzpatrick JA, León L, Arsuaga J-L, Falgueres JJB, Bullen T. 1997. Geology and preliminary dating of the hominid-bearing sedimentary fill of the Sima de los Huesos Chamber, Cueva Mayor of the Sierra de Atapuerca, Burgos, Spain. J Hum Evol 33: 129155.
  • Bischoff JL, Shamp DD, Aramburu A, Arsuaga JL, Carbonell E, Bermúdez de Castro JM. 2002. J Archaeol Sci 30: 275280.
  • Björk A. 1969. Prediction of mandibular growth. Am J Orthod 55: 585599.
  • Björk A. 1991. Facial growth rotation—reflections on definitions and cause. Proc Fin Dent Soc 87: 5158.
  • Bookstein FL. 1991. Morphometric tools for landmark data. Cambridge: Cambridge University Press. 455 p
  • Bookstein FL. 1996. Combining the tools of geometric morphometrics. In: MarcusLF, editor. Advances in morphometrics. New York: Plenum Press. p 131151.
  • Chernoff B, Magwene PM. 1999. Afterword. In: OlsonEC, MillerPL, editors. Morphological integration. Chicago: University of Chicago. p 319353.
  • Cheverud JM. 1982. Relationships among ontogenetic, static an evolutionary allometry. Am J Phys Anthropol 59: 139149.
  • Dryden IL, Mardia KV. 1998. Statistical shape analysis. Chichester: Wiley.
  • Emerson SB, Bramble DM. 1993. Scaling, allometry and skull design. In: HankenJ, HallBK, editors. The skull. Chicago: University of Chicago Press. p 384421.
  • Enlow DH. 1968. The human face: an account of the postnatal growth and development of the craniofacial skeleton. New York: Harper & Row. 423 p
  • Enlow DH, Hans MG. 1996. Essentials of facial growth. Philadelphia/London/New York: W.B. Saunders Company. 303 p.
  • Franciscus RG, Trinkaus E. 1995. Determinants of retromolar space presence in Pleistocene Homo mandibles. J Hum Evol 28: 577595.
  • Godfrey L, Sutherland M. 1996. Paradox of peramorphic paedomorphosis: heterochrony and human evolution. Am J Phys Anthropol 99: 1742.
  • Kendall DG. 1977. The diffusion of shape. Adv Appl Probabil 9: 428430.
  • Klingenberg CP. 1998. Heterochrony and allometry: the analysis of evolutionary change in ontogeny. Biol Rev 73: 79123.
  • Klingenberg CP, Leamy LJ. 2001. Quantitative genetics of geometric shape in the mouse mandible. Evolution 55: 23422352.
  • Klingenberg CP. 2002. Morphometrics and the role of the phenotype in studies of the evolution of developmental systems. Gene 287: 310.
  • Klingenberg CP, Mebus K, Auffray J-C. 2003. Developmental integration in a complex morphological structure: how distinct are the modules in the mouse mandible? Evol Dev 5: 522531.
  • Lieberman DE, McBratney BM, Krovitz G. 2002. The evolution and development of cranial form in Homo sapiens. Proc Natl Acad Sci U S A 99: 11341139.
  • Martinez-Maza C, Rosas A. 2002. Bone remodeling in the Atapuerca-SH mandibles. Implications for growth patterns in Middle Pleistocene hominids. Am J Phys Anthropol 115: 107108.
  • Maynard Smith J, Burian R, Kauffman S, Alberch P, Campbell J, Goodwin B, Lande R, Raup D, Wolpert L. 1985. Developmental constraints and evolution. Q Rev Biol 60: 265287.
  • Moss M, Young RW. 1960. A functional approach to craniology. Am J Phys Anthropol 45: 281292.
  • Penin X, Berge C, Baylac M. 2002. Ontogenetic study of the skull in modern humans and the common chimpanzees: neotenic hypothesis reconsidered with a tridimensional Procrustes analysis. Am J Phys Anthropol 118: 5062.
  • Ponce de León M, Zollikofer C. 2001. Neandertal cranial ontogeny and its implications for late hominid diversity. Nature 412: 534538.
  • Rohlf FJ, Slice D. 1990. Extensions of the Procrustes method for the optimal superimposition of landmarks. Syst Zool 39: 4059.
  • Rohlf FJ. 1996. Morphometric spaces, shape components and the effects of linear transformations. In: MarcusLF, editor. Advances in morphometrics. New York: Plenum Press. p 117128.
  • Rohlf FJ, Loy A, Corti M. 1996. Morphometric analysis of Old World talpidae (Mammalia, Insectivora) using partial warp scores. Syst Biol 45: 344362.
  • Rohlf F. 1998a. tpsREGR. version 1.27. Stony Brook NY: State University of New York.
  • Rohlf F. 1998b. SMALL. version 1.19. Stony Brook NY: State University of New York.
  • Rohlf F. 1998c. tpsRELW. version 1.30. Stony Brook NY: State University of New York.
  • Rosas A. 1992. Ontogenia y filogenia de la mandíbula en la evolución de los homínidos. Aplicación de un modelo de morfogénesis a las mandíbulas fósiles de Atapuerca. In. Madrid: Universidad Complutense.
  • Rosas A. 1995. Seventeen new mandibular specimens from the Atapuerca/Ibeas Middle Pleistocene hominids sample (1985–1992). J Hum Evol 28: 533559.
  • Rosas A. 1997. A gradient of size and shape for the Atapuerca sample and Middle Pleistocene hominid variability. J Hum Evol 33: 319331.
  • Rosas A. 1998. Modelos de crecimiento en mandíbulas fósiles de homínidos. Atapuerca, un nuevo paradigma. In: AguirreE, editor. Atapuerca y la Evolución Humana. Madrid: Fundación Ramón Areces. p 239275.
  • Rosas A. 2000. Ontogenic approach to variation in Middle Pleistocene hominids. Evidence from the Atapuerca-SH mandibles. Hum Evol 15: 8398.
  • Rosas A. 2001. Occurrence of Neandertal features in mandibles from the Atapuerca-SH Site. Am J Phys Anthropol 114: 7491.
  • Rosas A, Bastir M. 2002. Thin-plate spline analysis of allometry and sexual dimorphism in the human craniofacial complex. Am J Phys Anthropol 117: 236245.
  • Rosas A, Bastir M, Martínez Maza C, Bermúdez de Castro JM. 2002a. Sexual dimorphism in the Atapuerca-SH hominids. The evidence from the mandibles. J Hum Evol 42: 451474.
  • Rosas A, Bastir M, Martinez-Maza C. 2002b. Morphological integration and predictive value of the mandible in the craniofacial system of hominids: a test with the Atapuerca SH mandibular sample. Coll Antropol 26: 171172.
  • Sheets DH. 2001. IMP, integrated morphometric package.∼sheets/morphsoft.html; Sheets, David H.
  • Slice D. 2001. Landmark coordinates aligned by Procrustes analysis do not lie in Kendall's shape space. Syst Biol 50: 141149.
  • Smith KK. 1993. The form of the feeding apparatus in terrestrial vertebrates. In: HankenJ, HallBK, editors. The skull. Chicago/London: Chicago University Press. p 150196.
  • Zelditch ML, Lundrigan BL, David Sheets H, Garland T. 2003. Do precocial mammals develop at a faster rate? A comparison of rates of skull development in Sigmodon fulviventer and Mus musculus domesticus. J Evol Biol 16: 708720.