AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 118:324 –340 (2002) Brain Size and the Human Cranial Base: A Prenatal Perspective Nathan Jeffery* and Fred Spoor Evolutionary Anatomy Unit, Department of Anatomy and Developmental Biology, University College London, London WC1E 6JJ, UK KEY WORDS fetal ontogeny; basicranium; relative encephalization; MRI; human evolution ABSTRACT Pivotally positioned as the interface between the neurocranium and the face, the cranial base has long been recognized as a key area to our understanding of the origins of modern human skull form. Compared with other primates, modern humans have more coronally orientated petrous bones and a higher degree of basicranial flexion, resulting in a deeper and wider posterior cranial fossa. It has been argued that this derived condition results from a phylogenetic increase in the size of the brain and its subcomponents (infra- and supratentorial volumes) relative to corresponding lengths of the cranial base (posterior and anterior, respectively). The purpose of this study was to test such evolutionary hypotheses in a prenatal ontogenetic context. We measured the degree of basicranial flexion, petrous reorientation, base lengths, and endocranial volumes from high-resolution magnetic resonance images (hrMRI) of 46 human fetuses ranging from 10 –29 weeks of gestation. Bivariate comparisons with age revealed a number of temporal trends during the period investigated, most notable of which were coronal rotation of the petrous bones and basicranial retroflexion (flattening). Importantly, the results reveal significant increases of relative endocranial sizes across the sample, and the hypotheses therefore predict correlated variations of cranial base flexion and petrous orientation in accordance with these increases. Statistical analyses did not yield results as predicted by the hypotheses. Thus, the propositions that base flexion and petrous reorientation are due to increases of relative endocranial sizes were not corroborated by the findings of this study, at least for the period investigated. Am J Phys Anthropol 118: 324 –340, 2002. © 2002 Wiley-Liss, Inc. The cranial base of modern humans differs markedly from that of other primates (Fig. 1). The petrous pyramids are more coronally oriented, the foramen magnum faces more inferiorly, and it shows a higher degree of midline basicranial flexion, i.e., the basioccipital has a more vertically inclined orientation relative to the anterior cranial base (e.g., Keith, 1910; Duckworth, 1915; Zuckerman, 1955; Cameron, 1930; Ashton, 1957). Consequently, modern human crania have a deeper and wider posterior cranial fossa, than other primates, including the extant African apes (Dean, 1988; Ross and Ravosa, 1993; Spoor, 1997). Analysis of the evolutionary factors underlying this uniquely derived morphology is complex, owing to the cranial base’s central position in the head as the interface between the face, the neurocranium, and the neck. Based on specific structural demands that appear to follow from these relations a multitude of hypotheses have been proposed, which phylogenetically link the modern human morphology with, among others, brain expansion (i.e., phylogenetic encephalisation), obligatory bipedalism, and facial orthognathism (e.g., Biegert, 1963; Du Brul, 1977; Dean, 1988; Ross and Ravosa, 1993). Most recently, some of these hypothetical associations have been assessed by analyzing interspecific trends among extant and fossil primates (e.g., Ross and Ravosa, 1993; Ross and Henneberg, 1995; Spoor, 1997; Strait and Ross, 1999; Lieberman et al., 2000). Phylogeny can be seen as the product of successive ontogenies, as the accumulation of developmental change (Garstang, 1922). Indeed, a number of studies advanced possible ontogenetic mechanisms underlying the phylogenetic origin of the modern human cranial base (e.g., Ford, 1956; Moss, 1958; Scott, 1958; Knowles, 1963; Enlow and Hunter, 1968; Enlow, 1976). Such hypotheses generally take a mechanistic approach, in which the cranial base responds during ontogeny to the differential growth and development of closely related structures, such as the brain, or aspects of the face. These mechanis- © 2002 WILEY-LISS, INC. Grant sponsor: Medical Research Council; Grant sponsor: Wellcome Trust; Grant sponsor: University of London Intercollegiate NMR Research Service Scheme, Queen Mary and Westfied College. *Correspondence to: Nathan Jeffery, Evolutionary Anatomy Unit, Department of Anatomy and Developmental Biology, University College London, Rockefeller Building, University St., London WC1E 6JJ, UK. E-mail email@example.com Received 30 April 2001; accepted 11 October 2001. DOI 10.1002/ajpa.10040 Published online in Wiley InterScience (www.interscience.wiley. com). FETAL ENCEPHALIZATION AND BASE ARCHITECTURE Fig. 1. Sketches of midline and transverse sections through the primate skull. a: Loris. b: Chimpanzee. c: Modern human. Differences of basicranial flexion (top row) and petrous bone orientation (bottom row) are highlighted in grey. Notice that the cranial base is flatter and the petrous bones are more sagittally orientated in the chimpanzee than in the modern human skull. Not to scale. tic explanations are often contrasted with evidence of direct genetic control, emerging from studies of mutations and knockout gene experiments (reviews in Herring, 1993; Thorogood, 1993). A broader interpretation is that the ontogenetic basis for phylogenetic change contains both genetic and mechanical elements (e.g., Baker, 1941; Scott, 1954; Thorogood, 1987, 1988). While the genetic contributions to skull development and evolution were the focus of numerous studies (reviews in Sperber, 1992; Scheuerle, 1995; Thesleff, 1998), the mechanistic determinants remain largely unexplored by comparison. This study has two main aims. Firstly, it seeks to document major changes in basicranial and brain morphology during human fetal development, focusing on those aspects that uniquely characterize modern humans. Secondly, using this information it tests phylogenetic hypotheses describing the interaction between the brain and cranial base, in a first systematic attempt to evaluate possible ontogenetic mechanisms that may contribute to the derived cranial morphology of modern humans. PHYLOGENETIC PERSPECTIVE Now often referred to as spatial-packing hypotheses, phylogenetic scenarios linking the origin of the modern human cranial base with encephalization suggest that base flexion and petrous reorientation are solutions to a “spatial-packing” problem created by the phylogenetic increase of relative brain size (e.g., Cameron, 1924; Biegert, 1963; Gould, 1977; Dean, 1988; Ross and Ravosa, 1993; Lieberman et al., 2000). This implies that in modern humans, the basioccipital and foramen magnum were driven anteroinferiorly, the anterior cranial base was deflected ventrally, and the petrous bones reoriented coronally to accommodate a relatively larger brain. Inspired by Gould (1977), Ross and Ravosa (1993) formulated a specific version of the spatial-packing hypothesis, which proposes that the derived nature 325 of the modern human basicranium follows from the combination of a large brain and a short cranial base. The authors tested the hypothesis and established among extant primate species a correlation of greater cranial base flexion with increases of brain volume measured relative to midline cranial base length (i.e., index of relative encephalization 1, IRE1). Ross and Henneberg (1995) demonstrated that modern humans deviate from this primate trend by showing less flexion than expected for their relative brain sizes. This they interpret as evidence of a limit to the degree of flexion, perhaps imposed by structural demands other than relative encephalization (e.g., maintaining airway patency). Spoor (1997) also found an interspecific association between cranial base flexion and relative brain size, but in contrast to Ross and Henneberg (1995), showed that modern humans do not deviate significantly from the primate trend. This discrepancy between the studies of Spoor (1997) and Ross and Ravosa (1993) and Ross and Henneberg (1995) probably reflects differences in the landmarks used to measure cranial base angle and length (see McCarthy, 2001). In addition to examining basicranial flexion, Spoor (1997) also assessed the orientation of the foramen magnum and the petrous pyramids. He found that increases of relative brain size across extant primate species are concurrent with coronal reorientation of the petrous pyramids and a more inferior-facing foramen magnum. In his study, modern humans follow the primate trend for the orientation of the foramen magnum, but their petrous pyramids are less coronally orientated than predicted. As with base flexion, this could imply that petrous reorientation is in some way constrained. A recent study by Lieberman et al. (2000) adds considerable evidence in support of the hypothesis of Ross and Ravosa (1993). The authors found significant correlations between relative brain size and cranial base angle across extant primates while controlling for the influence of phylogenetic relationships among species studied (see Harvey and Pagel, 2000), using different measures of cranial base angle and relative brain size, as well as at two taxonomic levels (primates and haplorhines). Rather than relating the spatial-packing concept to entire brain size, Moss (1958) focused on the posterior cranial fossa. He noted that confinement of the modern human cerebellum to an inadequately sized posterior fossa, due to skull deformation, is invariably accompanied by increased flexion of the cranial base. Later, Dean (1988) proposed that the derived modern human basicranium follows from an enlarged cerebellum combined with a short posterior cranial base. That is, enlargement of the posterior cranial fossa, through coronal reorientation of the petrous pyramids and increased basicranial flexion, is a structural solution to the spatial-packing problem which followed from disproportionate enlargement of the cerebellum relative to the posterior fossa length. Ross and Ravosa (1993) tested this hypoth- 326 N. JEFFERY AND F. SPOOR esis by comparing absolute cerebellar volume against cranial base angle for an interspecific sample of extant primates. They found no correlation, but this does not refute the hypothesis put forward by Dean (1988), as the latter pertains to relative rather than absolute cerebellar size. A different perspective on the spatial-packing idea is found in hypotheses which link base morphology with size increases in one part of the brain relative to another. Hofer (1969) was among the first to propose such a hypothesis. He suggested that differential enlargement of the cerebrum not only maximizes its sphericity, and thereby provides the potential adaptive advantages of reduced neural wiring lengths and better skull balance (see Jerison, 1982; Ross and Henneberg, 1995), but also necessitates a flexion between the cerebrum and brain-stem that is matched by flexion of the midline basicranium. Similar explanations focusing on brain shape are posited elsewhere, and broadly speaking suggest that differential enlargement of the supratentorial brain (cerebrum and diencephalon) compared with the infratentorial brain (cerebellum and brain-stem) creates a “spatial-packing” problem with various predicted consequences for base architecture (e.g., Dean and Wood, 1981; Dean, 1988; Strait, 1999). In particular, Strait (1999) noted that anatomically different parts of the brain exhibit distinct scaling trajectories across extant primate species, and he hypothesized that phylogenetic increases of neocortical (cerebral neocortex) relative to noncortical (medulla, mesencephalon, and diencephalon) parts of the brain might be responsible for the derived condition of the modern human basicranium. A recent interspecific evaluation of the influence of differential encephalization revealed a significant correlation between increases of cerebral volume over brainstem volume and cranial base flexion across extant primates (Lieberman et al., 2000). ONTOGENETIC PERSPECTIVE The developmental literature contains several hypotheses, ideas, and findings that are congruous with the evolutionary mechanisms described above. Perhaps most striking are the similarities between the hypothesis of Ross and Ravosa (1993), formulated to explain flexion during human evolution, and a model Enlow (1976) detailed to explain flexion during human growth. Enlow (1976) proposed that cranial base flexion during human prenatal and postnatal ontogeny is due to increased brain growth relative to slower growth of the midline basicranium. This is consistent with human fetal studies showing that growth along the cranial base is significantly slower than in other parts of the skull (e.g., Houpt, 1970; Mandarim-de-Lacerda and Alves, 1992; Plavcan and German, 1995) and that fetal brain growth is markedly rapid at the same time (e.g., Jenkins, 1921; Koop et al., 1986). Taken together, these findings imply that relative brain size increases during human fetal life. Previous investi- gations also demonstrated simultaneous changes in fetal base morphology, particularly in base angulation (e.g., Burdi, 1965; Cousin, 1969; Erdoglija, 1989, 1990; Dimitriadis et al., 1995), but these are difficult to interpret due to inconsistencies in the morphological landmarks and methods of visualization these studies used. For example, Erdoglija (1989) and Dimitriadis et al. (1995) employed x-raybased imaging techniques, i.e., plain film radiography and computed tomography, to visualize the fetal cranial base. These modalities rely on the radiodensity of tissues to highlight morphologies. While these techniques are suitable for imaging radiodense structures like bone, they are not effective in imaging the incompletely ossified structures of the fetal cranial base, where key landmarks are often defined by less radiodense tissues like cartilage. In contrast to these x-ray-based modalities, the technique used in the present study, known as magnetic resonance imaging (MRI), relies on the quantum properties of hydrogen nuclei and allows for the noninvasive visualization and differentiation of many more soft tissues, as well giving indirect views of ossified bone (see Spoor et al., 2000a,b). Several developmental studies allude to an ontogenetic process of relative cerebellar enlargement, similar to the process Dean (1988) described as an evolutionary mechanism. Ford (1956) suggested that fetal changes of base morphology are a compensatory mechanism for slower growth of the posterior cranial base compared with that of the anterior cranial base and brain. Fetal studies confirm that growth of the posterior cranial base is slower than that of the anterior base (Burdi, 1965, 1969; Levihn, 1967; Houpt, 1970). Moreover, following a slow start, fetal cerebellar growth accelerates with sizable increases of surface area, mass, and width at rates exceeding those of the cerebrum (Noback and Moss, 1956; Dobbing and Sands, 1973; Rakic and Sidman, 1970). The likely outcome of this growth spurt, combined with conservative growth along the posterior cranial base, is relative infratentorial enlargement within the confines of the posterior cranial fossa. Findings from fetal studies are also consistent with processes of ontogenetic differential encephalization. Marked increases in the supratentorial portion of the brain occur in relation to the infratentorial portion during fetal development (GuihardCosta and Larroche, 1990, 1992). More importantly, Moss et al. (1956) linked fetal differential encephalisation with changes of posterior fossa morphology. The authors noted that critical developmental phases in growth of the occipital and temporal bones correlate with changes in overall proportions of the fetal brain. HYPOTHESES Based on the reviewed models of phylogenetic and ontogenetic interaction between the brain and cranial base, the following hypotheses were tested for FETAL ENCEPHALIZATION AND BASE ARCHITECTURE morphological change during the second and part of the third trimester of human fetal development. General spatial-packing The first hypothesis predicts that increases of fetal brain volume relative to cranial base length create a spatial-packing problem that results in base flexion and coronal petrous reorientation. Three of the possible outcomes are: 1) variations of cranial base angle and petrous orientation correlate with changes of brain volume relative to base length and thus, the hypothesis is corroborated; 2) cranial base angle and petrous orientation remain essentially unchanged if cranial base length scales isometrically with brain volume, i.e., relative brain size remains unchanged and the hypothesis is not applicable; and 3) there are no correlations or isometric scaling, and the hypotheses is uncorroborated. The model of spatial-packing following Ross and Ravosa (1993) (cube root brain volume/base length) assumes that the midline cranial base is influenced by expansion of the brain in all directions. However, it could be that relative brain size increases mostly due to, for example, lateral expansion of the brain without impinging on the two-dimensional midline basicranium. Hence, it should be assessed whether increases of the midline endocranial area, which is spatially associated with the midline cranial base, are equal to or greater than overall increases of brain volume. Infratentorial spatial-packing This hypothesis predicts that increases of infratentorial volume measured against the length of the posterior cranial base create a “spatial-packing” problem within the posterior cranial fossa that results in cranial base flexion and coronal petrous reorientation. The three possible outcomes are essentially the same as those described for the general spatial-packing hypothesis. Differential encephalization The third hypothesis predicts that prenatal increases of supratentorial volume relative to infratentorial volume impinge on the cerebellum and brain-stem, creating a spatial-packing problem that drives coronal reorientation of the petrous bones and cranial base flexion. Possible outcomes include: 1) differential encephalization in association with base flexion and coronal petrous reorientation; 2) no differential encephalization, and consequently the hypothesis is not testable; and 3) differential encephalization but no correlation with either base flexion or coronal petrous reorientation, and hence the hypothesis is uncorroborated. MATERIALS AND METHODS Sample Fifty formalin-fixed human fetuses from museum collections of therapeutic and natural abortuses 327 housed at the Department of Anatomy and Developmental Biology, University College London, and the Department of Anatomy, Queen Mary and Westfield College, London, were imaged with high-resolution magnetic resonance imaging (hrMRI). Although specimens were carefully screened for pathologies before scanning, four individuals were excluded from the study because of abnormalities shown in the images. Estimated gestational age (EGA) was computed to the nearest tenth of a week for each fetus from measures of biparietal diameter with reference to the regression equation for exocranial measures given in Chitty et al. (1994). Ages were calculated to the nearest tenth of a week to maintain the distinction between individuals provided by biparietal diameter while presenting the data in a more readily interpreted format, i.e., that of gestational time. Other, perhaps more suitable age proxies (see Sherwood et al., 2000) were precluded due to previous invasive work on the material. The age range of the specimens thus calculated was 10 –29 weeks of gestation (Table 1). The sample distribution in respect of biparietal diameter was neither significantly skewed nor kurtotic (skew ⫽ ⫺0.08, ns; kurtosis ⫽ ⫺0.25, ns). The fetuses were imaged using a 4.7T imaging and spectroscopy unit (Sisco-Varian). The following parameters and pulse sequences were used: 10 Gcm⫺1 (max) gradient coil; an 8.8-cm (internal diameter) saddle coil; and a 6.2-cm birdcage coil, according to the size of the specimen; matrix size, 256 ⫻ 256; number of averages (NEX), 24 with typical scanning times of 18 –22 hr; and a T2-weighted spin-echo multislice sequence (TR and TE were approximately 9,000 msec and 40 msec, respectively). Fields of view (FOV) were adjusted according to the size of the specimen, and varied from 30 –115.2 mm, giving a pixel size of 0.12– 0.45 mm. The T2 weighting was employed in preference to T1 in order to maximize contrast between the formalin solution (high water content) surrounding the brain and the skull base. Images were acquired contiguously along the transverse (axial) plane, with a slice thickness of between 0.23– 0.90mm. Coordinates for anatomical landmarks were taken from the original transverse images and from resampled sagittal and coronal images (AVS5, Advanced Visual Systems). Slice thickness, field of view, biparietal diameter, and computed age are given in Table 1 for each specimen. Pixel sizes for each specimen can be computed by dividing the field of view (FOV) by the acquisition matrix size (i.e., 256). Measurements The degree of basicranial flexion was measured as the cranial base angle (CBA) computed from landmark coordinates of the foramen caecum (Fc), sella (S), and basion (Ba), taken in midsagittal images. Some developmental studies used nasion rather than foramen caecum to measure CBA (e.g., Ford, 1956; Burdi, 1965). However, the nasion is a facial 328 N. JEFFERY AND F. SPOOR 1 TABLE 1. Sample Code BPD (mm) EGA (weeks) FOV (mm) Slice thickness (mm) UMSCL-163 UMSCL-39 QMWF49 UMSCL-61 QMWF7 UMSCL-2992 QMWF34 UMSCL-164 UMSCL-962 011A2 UMSCL-165 UMSCL-94 Z101 UMSCL-63 UMSCL-92 UMSCL-216 QMW35 QMW28 UMSCL-223 UMSCL-86 UMSCL-91 UMSCL-287 Z102 UMSCL-220 UMSCL-2222 UMSCL-221 UMSCL-210 ZN1 UMSCL-166 UMSCL-215 UMSCL-212 UMSCL-207 UMSCL-218 UMSCL-214 UMSCL-211 UMSCL-297 UMSCL-209 UMSCL-292 UMSCL-286 UMSCL-281 UMSCL-301 UMSCL-284 UMSCL-242 UMSCL-248 UMSCL-232 UMSCL-229 UMSCL-238 UMSCL-275 UMSCL-240 UMSCL-303 15.0 15.0 19.0 21.0 24.0 24.0 25.5 27.0 29.0 32.0 32.0 33.0 35.0 37.0 37.0 38.5 40.0 40.0 40.0 42.0 42.0 42.0 43.0 43.0 44.0 44.0 45.0 45.0 46.0 46.0 47.0 48.0 49.0 49.0 49.0 50.0 52.0 52.0 54.0 57.0 57.5 60.5 62.0 64.0 64.0 66.0 67.0 69.5 70.0 77.0 10.1 10.1 11.5 12.1 13.1 13.2 13.5 14.0 14.1 15.4 15.4 15.7 16.3 16.8 16.8 17.2 17.6 17.6 17.6 18.2 18.2 18.2 18.4 18.4 18.7 18.7 19.0 19.0 19.2 19.2 19.5 19.8 20.1 20.1 20.1 20.3 20.9 20.9 21.5 22.3 22.5 23.4 23.8 24.5 24.5 25.1 25.5 26.3 26.5 29.2 40 30 40 40 40 40 40 50 50 50 64 50 64 64 64 64 64 80 64 64 64 80 64 80 64 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 102.4 85 89.6 102.4 102.4 102.4 102.4 115.2 0.31 0.23 0.31 0.31 0.31 0.31 0.31 0.39 0.39 0.39 0.50 0.39 0.50 0.50 0.50 0.50 0.50 0.63 0.50 0.50 0.50 0.63 0.50 0.63 0.50 0.64 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.80 0.66 0.70 0.80 0.80 0.80 0.80 0.90 1 BPD, biparietal diameter; EGA, estimated gestation age; FOV, imaging field of view. Specimens with codes commencing with UMSCL are from University College London; codes starting with Z, QMW, or 01 refer to those specimens from Queen Mary and Westfield College, London. 2 Excluded from study. rather than a basicranial landmark, and its use is inappropriate because it introduces variation from facial growth (Enlow and Moyers, 1971). Ideally, measures of CBA would attempt to approximate the primary points of flexion, i.e., the spheno-ethmoidal, midsphenoidal, and spheno-occipital synchondroses (see Lieberman and McCarthy, 1999). However, not all of the synchondroses were sufficiently resolved with hrMRI throughout the period investigated. The landmarks used were defined as follows (Fig. 2a,b; Table 2): Fc, the center of the endocranial opening of Fig. 2. High-resolution magnetic resonance images (hrMRI) of fetal head, and schematic representations of measures taken thereon. a: Midline hrMR image of fetal head. b: Sketch showing measure of cranial base angle computed from coordinates for foramen caecum (Fc), sella (S), and basion (Ba). c: Transverse hrMR image at level of anterior semicircular canals and presphenoid. d: Sketch showing measure of interpetrosal angle computed from coordinates for medialmost (Md) and lateralmost (Lt) attachments of tentorium cerebelli to petrous ridges. Scale bars, 10 mm. the foramen caecum, seen as a pit on the cribriform plate between the fetal crista galli and the endocranial wall of the frontal bone; S, the centerpoint of the sella turcica; and Ba, the tip or point of greatest curvature of the basioccipital on the anterior margin of the foramen magnum. A number of studies measured petrous orientation using exocranial landmarks (e.g., Dean and Wood, 1981, 1984) and landmarks on the endocranial petrosal surface (Putz, 1974; Spoor, 1997; Spoor and Zonneveld, 1995, 1998), but these cannot be replicated in the developing fetus. Here, petrous orientation was approximated with a measure of interpetrosal angle (IPA). This was defined as the angle between line segments fitted through the lateralmost to the medialmost attachments of the tentorium cerebelli to the petrous ridges (Fig. 2c,d; Table 2). In practice, landmarks were taken from coronal images anterior to the cochleae and just anterior to the sigmoid sulci. Total, anterior, and posterior cranial base lengths were measured as the distances between Fc, S, and Ba (Fig. 2b; Table 2). Thus, unlike the definition in FETAL ENCEPHALIZATION AND BASE ARCHITECTURE 329 TABLE 2. Landmarks, angles, and volume measurements Measure Abbreviation Landmarks Basion Foramen caecum Ba Fc Sella Medial petrous S Md Lateral petrous Lt Linear measures Total cranial base length Anterior cranial base length Posterior cranial base length Angles Cranial base angle Interpetrosal angle Volumes and areas Endocranial volume Supratentorial volume Infratentorial volume Midline area Derived measures Relative endocranial size TBL ABL PBL CBA IPA Definition Midline point on anterior margin of foramen magnum Midline point marking pit between fetal crista galli and endocranial wall of frontal bone Center of sella turcica in midline Points marking medialmost tentorial attachments to left and right petrous ridges Points marking lateralmost tentorial attachments to left and right petrous ridges Total linear distance from basion (Ba) to sella (S), and sella to foramen caecum (Fc) Linear distance from sella (S) to foramen caecum (Fc) Linear distance from basion (Ba) to sella (S) Ventral angle between midline anterior cranial base (Fc-S) and posterior cranial base (S-Ba) Posterior angle between left and right line segments fitted through medialmost (Md) and lateralmost (Mt) tentorial attachments to petrous ridges EV SV IV MA Sum Sum Sum Sum IRE 1 Cube root of endocranial volume divided by total cranial base length (EV0.33/ TBL) Cube root of infratentorial volume divided by posterior cranial base length (IV0.33/PBL) Sum of endocranial voxels below tentorium cerebelli, divided by sum of voxels above tentorium (IV/SV) Relative infratentorial size RIE Index of differential encephalisation IDE of of of of voxels within endocranial cavity voxels within endocranial cavity above tentorium cerebelli voxels within endocranial cavity below tentorium cerebelli pixels representing midline plane of endocranial cavity Ross and Ravosa (1993), the anterior cranial base length Fc–S includes the cribriform plate. This is because the ethmoid region is ontogenetically recognized as part of the cartilaginous base plate and thus of the cranial base (De Beer, 1937; Shapiro and Robinson, 1980). Moreover, the cribriform plate forms a substantial portion of the endocranial surface available to the developing brain, and omission in tests of spatial packing hypotheses would not be justified (see discussion in McCarthy, 2001). The midsagittal endocranial area was measured as the number of pixels in the outlined region of interest multiplied by pixel area. Endocranial, infratentorial, and supratentorial volumes were measured by manually outlining regions in all sagittal images showing the anatomy of interest (typically 70 –90 images), taking care not to include normal dural structures such as the confluence of sinuses, falx cerebelli, or tentorium cerebelli. Regions of interest were outlined using the trace function in AVS5 (Advanced Visual Systems). For each slice, the number of pixels in the region of interest was multiplied by pixel area (i.e., 0.014 – 0.203 mm2) and subsequently by slice thickness (i.e., 0.23– 0.90 mm). This gave a volume of interest for each slice, and the sum for all relevant slices provided a measure of volume. Tests of accuracy indicate that this method incurs an error of between 1–3% (Jeffery, 1999). Infratentorial volumes were measured using the same method by outlining the endocranial region below the tentorium cerebelli. Supratentorial volumes were calculated by subtracting infratentorial volume from endocranial volume. All measurements were taken by the same rater (N.J.) and are available upon request. Statistical analyses Measurements for each individual were computed to the nearest pixel size rounded to tenths of a millimeter. For example, measures for individuals imaged with a 40 ⫻ 40 mm and 50 ⫻ 50 mm FOV were taken to the nearest two tenths of a millimeter, because their pixel size was 40/256 ⫽ 0.156 and 50/256 ⫽ 0.195, respectively. Angular measurements were taken to the nearest degree, area measurements were taken to the nearest square millimeter, and volume measurements were taken to the nearest 10 cubic millimeters. Volumes and areas were reduced for linear comparisons using cube and square roots, respectively. Relative endocranial size (IRE 1) was computed as the cube root of endocranial volume over cranial base length. Relative infratentorial size (RIE) was calculated as the cube root of infratentorial volume over posterior base length, and the index of differential encephalization (IDE) was calculated as the ratio of infratentorial volume over supratentorial volume. Measurement precision was evaluated by comparison of five repetitions for four randomly selected fetuses over 5 consecutive days. One-way analysis of variance (ANOVA) of repeated measurements (see Table 3) showed that errors incurred among repeated measurements are negligible in comparison with the biological variation between individual fetuses. In 330 N. JEFFERY AND F. SPOOR TABLE 3. Results for one-way ANOVA test of linear, angular, and volume measurements showing mean square differences between fetuses and repeated measures1 Measurement MS repeated measures MS between fetuses F P2 Anterior cranial base length Posterior cranial base length Total cranial base length Cranial base angle Interpetrosal angle Endocranial volume (⫻10⫺3) Supratentorial volume (⫻10⫺3) Infratentorial volume (⫻10⫺3) Midline area (⫻10⫺2) 0.125 0.022 0.105 1.851 24.174 0.608 0.479 0.057 4.061 52.167 20.574 160.467 49.198 150.300 3,265.894 2,669.377 30.148 14,458.750 417.333 924.674 1,531.908 26.579 6.217 5,367.342 5,571.894 524.501 35.606 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.01 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 1 2 MS, mean square. n ⫽ 20, k ⫽ 5; F0.001(3,16) ⫽ 9.0, F0.01(3,16) ⫽ 5.29, F0.05(3,16) ⫽ 3.24. other words, the null hypothesis that values for repeated measures from one individual fetus are the same was accepted (P ⬍ 0.01). The relationships between variables were evaluated by nonparametric methods and without logarithmic transformations. Spearman rank correlation coefficients (rrank) and model II reduced majoraxis (RMA) regression equations were employed to investigate bivariate trends between variables (Sokal and Rohlf, 1995). Complex trends were described with second-order polynomial regressions and, in one case, a two-step RMA. Partial productmoment correlations were used to evaluate the effects of colinearities with age, nee biparietal diameter, on rank correlations used to test the hypotheses. t-tests were used to test the statistical significance of correlation and RMA statistics. In all significance tests, a level of P ⬍ 0.05 was used to reject null hypotheses. RESULTS Temporal changes Mean values of the linear, angular, and volumetric measures, computed at 2-week intervals of EGA, are listed in Table 4 to illustrate the magnitude of their variations during the 10 –29-week period represented by the sample. Angular measures. Figure 3a plots CBA and IPA against EGA. This reveals a significant positive correlation for both variables with age (Table 5). Over the period investigated, the cranial base is shown to retroflex, i.e., flatten out, by some 9°, and the petrous bones are shown to reorientate coronally, i.e., away from one another, by 10 –15°. Length measures. Plots of total, posterior, and anterior cranial base lengths against EGA show notable increases (Fig. 3b). The slope of anterior base length against EGA was almost twice that of the posterior base (Table 5). This demonstrates that increases of total cranial base length are mostly due to elongation of the anterior cranial base. Volumetric measures. Comparisons of absolute endocranial volume against age yield a second-order polynomial (i.e., parabola) fit of y ⫽ 664.75x2 ⫺ 14,686x ⫹ 84,264 (R2 ⫽ 0.966). The enlargement was markedly rapid, and described a 136-times increase from an initial mean value of 1,625 mm3 at 10.1 weeks EGA to 220,520 mm3 at 29.2 weeks EGA. The endocranial volume is compartmentalized by the tentorium cerebelli into infratentorial and supratentorial components. Temporal variations of these volumes against EGA were described with a second-order polynomial of y ⫽ 39.877x2 ⫺ 757.21x ⫹ 4037.9 (R2 ⫽ 0.930) for infratentorial volume, and of y ⫽ 625.91x2 ⫺ 13969x ⫹ 80563 (R2 ⫽ 0.966) for supratentorial volume. These represent 48- and 158-fold increases of their initial mean values at 10.1 weeks EGA, respectively. In order to clarify these trends, volumes were linearized using cube roots, i.e., volume1/3 and plotted against EGA (Fig. 3c). The slope of cube root supratentorial volume against EGA is over twice that of cube root infratentorial volume against EGA (Table 5). These plots show that increases of overall endocranial volume are mostly due to increases of supratentorial volume as opposed to infratentorial volume. Note that anterior cranial base length is spatially associated with supratentorial volume, as is posterior cranial base with infratentorial volume. Derived ratios. Temporal changes of relative brain sizes and IDE were evaluated with comparisons against EGA. A significant linear increase of relative endocranial size was observed with EGA (Table 5). However, no significant change in relative infratentorial size was revealed in relation to EGA (rrank ⫽ 0.084, ns). Bivariate comparisons of IDE with EGA produced a negative polynomial curve fitting of y ⫽ 0.0008x2 ⫺ 0.0388x ⫹ 0.5488 (R2 ⫽ 0.817). This describes an early period of rapid differential enlargement in which supratentorial volume increases in relation to infratentorial volume, i.e., the ratio of infratentorial volume/supratentorial volume decreases (Fig. 3d). This trajectory gradually tapers off towards week 20 of EGA, and from weeks 20 –29 it levels almost to plateau. The variation in IDE before week 20 of EGA ranges from 0.07– 0.27, whereas that after week 20 of EGA ranges from 0.06 – 0.11. BPD, biparietal diameter; ABL, anterior base length; PBL, posterior base length; TBL, total base length; CBA, cranial base angle; IPA, interpetrosal angle; EV, endocranial volume; IV, infratentorial volume; SV, supratentorial volume; MA, midline endocranial area; s.d., one standard deviation; wks, weeks. 1 141 134 19 52 129 117 190 227 212 277 520 886 1,121 1,678 1,998 2,456 3,110 4,328 1,663 1,931 1,381 2,637 3,815 2,979 4,055 6,008 5,597 2,733 9,333 18,468 29,071 47,431 66,843 92,440 126,523 192,960 273 382 221 266 272 435 581 433 416 558 1,348 2,580 2,898 4,459 5,913 8,050 10,453 15,050 1,933 2,306 1,558 2,860 4,069 3,303 4,555 6,297 5,748 3,290 10,680 21,048 31,969 56,548 72,757 100,490 136,975 208,010 74.3 77.0 76.0 75.8 79.3 87.7 83.3 81.8 91.5 6.4 9.3 11.2 12.3 14.1 15.5 16.4 18.1 18.7 10–13 13–15 15–17 17–19 19–21 21–23 23–25 25–27 27–29 wks wks wks wks wks wks wks wks wks 4 4 4 9 12 3 4 4 2 15–21 24–32 33–37 38.5–44 43–52 52–57 57.5–64 64–70 69.5–77 8.2 12.5 16.4 18.2 21.7 24.1 24.9 26.1 31.0 1.5 1.8 0.8 1.3 2.1 1.8 1.9 1.5 1.4 1.2 1.2 0.7 1.0 1.0 1.0 0.4 0.6 0.8 14.6 21.8 27.5 30.5 35.8 39.6 41.3 44.2 49.6 2.7 3.0 1.5 2.3 3.1 2.8 2.3 2.0 1.9 132.0 133.3 136.5 138.0 134.2 134.0 140.3 139.8 141.5 4.1 3.7 2.2 3.1 2.4 3.0 1.9 1.7 2.1 3.2 5.8 5.9 9.2 8.8 7.1 6.6 3.4 3.7 s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean — N Age EV (mm3) IPA (°) CBA (°) TBL (mm) PBL (mm) ABL (mm) BPD (mm) TABLE 4. Descriptive statistics for fetal measurements at 2-week intervals1 IV (mm3) SV (mm3) MA (mm2) FETAL ENCEPHALIZATION AND BASE ARCHITECTURE 331 General spatial-packing In order to test the general model of spatial-packing (cube root volume/base length), a comparison of square root midline endocranial area against cube root of endocranial volume was made. This demonstrates that increases of the midline area are larger than the overall increase in brain volume, and that lateral brain expansion is not predominant (Table 5; slope ⬎1 at P ⬍ 0.001). Increases of relative endocranial size across the sample were examined with a comparison of cube root endocranial volume against total base length (Table 5). The slope is significantly greater than 1 (P ⬍ 0.001). This shows that relative endocranial size increases significantly across the sample. The general spatial-packing hypothesis predicts that this increase of relative endocranial size will correspond with a decrease of CBA and an increase of IPA. However, the opposite relationship was found for CBA, which shows a small, statistically significant correlation with relative endocranial size that is positive rather than negative (Fig. 4a, Table 5). As predicted, increases of interpetrosal angle are positively correlated to increases of relative endocranial size, though only weakly (Fig. 4b, Table 5). Partial correlation coefficients for CBA and IPA against relative endocranial size are insignificant, while holding biparietal diameter constant (r ⫽ ⫺0.017, ns; r ⫽ ⫺0.064, ns). Infratentorial spatial-packing To evaluate variations of relative infratentorial size across the sample, the cube root of infratentorial volume was compared against posterior base length. This produces a slope that is significantly greater than 1 (Table 5) and shows that relative infratentorial size increased across the sample. The hypothesis predicts a correlated decrease of CBA and increase of IPA, but neither of these correlations was significant (rrank ⫽ 0.119, ns, and 0.008, ns, respectively). Thus, increases of infratentorial volume relative to posterior base length are not accommodated in the posterior cranial fossa by base flexion or coronal reorientation of the petrous bone. Differential encephalization Comparisons of supratentorial and infratentorial volume against EGA, as well as IDE against EGA, reveal differential enlargement in which the supratentorial volume increases disproportionately (Fig. 3c,d). The differential encephalization hypothesis therefore predicts a correlated increase of IPA and a correlated decrease of CBA. Comparisons of CBA and IPA against IDE revealed small significant negative correlations (rrank ⫽ ⫺0.297, P ⬍ 0.05, and ⫺0.481, P ⬍ 0.001, respectively). However, as Figure 5 demonstrates, these trends are clearly biphasic, not linear. In the first phase, there is little variation of the index of differential encephalization 332 N. JEFFERY AND F. SPOOR Fig. 3. Bivariate plots of measured variables against estimated gestational age (EGA). a: Cranial base angle (CBA) and interpetrosal angle (IPA) against EGA. b: Total base length (TBL), anterior base length (ABL), and posterior base length (PBL) against EGA. c: Cube root endocranial volume (EV), supratentorial volume (SV), and infratentorial volume (IV) against EGA, with reduced major axis (mra) line fittings shown. d: Index of differential encephalisation (IDE) against EGA, with the second-order polynomial line fitting shown. with either CBA or IPA. This is followed by a second phase in which the index of differential encephalization increases, but there is little change in either CBA or IPA. Indeed, the phases are so distinct that neither complete trend was successfully resolved with second-order polynomial fittings (R2 ⬍ 0.2). Note from Figure 3d that the first phase revealed in Figure 5 corresponds to later fetal life, and that the second phase corresponds to early fetal life. To explore these phases pragmatically, these data were subdivided for each ordinate variable according to mean abscissa value (Fig. 5). Each of the four subsets, two for both IPA and CBA, reflected their respective phases well and were analyzed separately. These analyses revealed that CBA is not significantly correlated to IDE for either phase one (IDE values ⬍ IDEmean; rrank ⫽ 0.055, ns) or two (IDE values ⬎ IDEmean; rrank ⫽ 0.113, ns). Furthermore, interpetrosal angle was not significantly correlated to IDE for the second phase (rrank ⫽ 0.393, ns). However, analysis of IPA against IDE for the first phase yields a significant negative correlation (Fig. 5; Table 5). These subanalyses show that CBA is independent of differential encephalization. It is also suggested that coronal reorientation of the petrous bones is associated with disproportionate supratentorial enlargement, but only as the process of differential encephalization tapers off to a state of equilibrium later in fetal life. However, the partial correlation coefficient for IPA against IDE for the first phase is insignificant when biparietal diameter is held constant (r ⫽ ⫺0.196, ns). In addition, partial correlation coefficients for both IPA and CBA against IDE are insignificant for the whole sample when biparietal diameter is held constant (r ⫽ 0.178, ns; r ⫽ 0.154, ns). FETAL ENCEPHALIZATION AND BASE ARCHITECTURE 333 1 TABLE 5. Reduced major axes and rank correlation statistics Plot rrank P Slope 95% confidence interval of slope Intercept CBA ⫻ EGA IPA ⫻ EGA TBL ⫻ EGA ABL ⫻ EGA PBL ⫻ EGA EV1/3 ⫻ EGA SV1/3 ⫻ EGA IV1/3 ⫻ EGA IRE ⫻ EGA MA1/2 ⫻ EV1/3 EV1/3 ⫻ TBL CBA ⫻ IRE IPA ⫻ IRE IV1/3 ⫻ PBL IPA ⫻ IDE, first phase 0.431 0.504 0.976 0.975 0.960 0.982 0.983 0.965 0.813 0.987 0.987 0.425 0.356 0.952 ⫺0.349 ** *** *** *** *** *** *** *** *** *** *** ** * *** * 0.98 1.60 2.10 1.33 0.78 2.65 2.62 0.98 0.02 1.10 1.26 60.79 99.43 1.25 ⫺557.51 0.72 ⬎ 1.24 1.18 ⬎ 2.02 1.96 ⬎ 2.37 1.23 ⬎ 1.43 0.72 ⬎ 0.84 2.53 ⬎ 2.78 2.50 ⬎ 2.75 0.91 ⬎ 1.04 0.01 ⬎ 0.02 1.06 ⬎ 1.14 1.20 ⬎ 1.33 43.80 ⬎ 77.78 71.57 ⬎ 127.28 1.16 ⬎ 1.34 ⫺749.64 ⬎ 365.39 117.49 48.75 ⫺7.03 ⫺5.61 ⫺1.72 ⫺15.15 ⫺15.69 ⫺2.73 0.75 ⫺0.99 ⫺6.23 71.63 ⫺26.26 ⫺0.60 131.97 1 CBA, cranial base angle; EGA, estimated gestational age; IPA, interpetrosal angle; TBL, total base length; ABL, anterior base length; PBL, posterior base length; EV, endocranial volume; SV, supratentorial volume; IV, infratentorial volume; IRE, relative endocranial size; MA, midline endocranial area; IDE, index of differential encephalisation. * P ⬍ 0.05. ** P ⬍ 0.01. *** P ⬍ 0.001. DISCUSSION Insight into the way morphological differences between closely related species emerge during ontogenetic development can make a major contribution to our understanding of the phylogenetic origins of the modern human cranium. The unique morphology among primates of the human cranial base has long been seen as the direct evolutionary consequence of encephalization (e.g., Virchow, 1857; Cameron, 1924; Dabelow, 1931; Biegert, 1957; Ross and Ravosa, 1993; Ross and Henneberg, 1995; Spoor, 1997; Lieberman et al., 2000; McCarthy, 2001). This study set out to assess possible ontogenetic mechanisms underlying this phylogenetic relationship by testing a number of specific hypotheses describing the interaction between the brain and cranial base. Temporal trends Cranial base angle. Findings reported here show cranial base retroflexion of about 9° in total during the second and early third trimesters. This is in agreement with a number of studies documenting basicranial retroflexion (e.g., Kvinnsland, 1971; Dimitriadis et al., 1995), and contrasts with studies suggesting stasis (e.g., Burdi, 1965, 1969) or basicranial flexion (e.g., Levihn, 1967; Erdoglija, 1989; van den Eynde et al., 1992) during development. These past studies often used small sample sizes (e.g., N⫽ 24: Burdi, 1965), inappropriate basicranial landmarks (e.g., nasion: van den Eynde et al., 1992), and/or x-ray-based methods of visualization with which radiotranslucent soft tissues are not clearly discerned (e.g., film radiography: Erdoglija, 1989). In contrast, the present study used a comparatively large sample (N ⫽ 46), carefully selected landmarks that were both spatially and temporally homologous (i.e., can be clearly defined in all individuals of the sample irrespective of developmental stage), and a method of visualization (MRI) that provided details of soft tissues as well as cartilage and bone. The findings of the present study are therefore more likely to be correct. Lieberman and McCarthy (1999) showed that the base flexes from 143° at 1 postnatal month to 134° at age 17 years, with the majority of flexion occurring within the first postnatal year. The authors used the same measure of cranial base angle as used in the present study. Together with the findings detailed in this study, the report of postnatal flexion contradicts the “fetalization” concept which proposes that the degree of base flexion is established during early fetal life and is subsequently retained into adulthood (see Zuckerman, 1955; Schultz, 1955). The aggregate of findings suggests that cranial base angulation is temporally dynamic, i.e., polyphasic, not fixed. Studies reveal phases of embryonic flexion (e.g., Müller and O’Rahilly, 1980; Diewert, 1983, 1985; Sperber, 2001), fetal retroflexion (e.g., Kvinnsland, 1971; Dimitriadis et al., 1995), and postnatal flexion (e.g., Lieberman and McCarthy, 1999). Interestingly, Lieberman and McCarthy (1999) reported an adult mean cranial base angle that is similar to the mean values reported here for fetal periods 10 –12 weeks and 13–15 weeks. It has already been shown that these similarities are not due to retention of the fetal condition. This suggests that temporally isolated, perhaps different, ontogenetic mechanisms or controls can independently afford resemblances, or convergence, between fetal and adult morphologies. The pattern of base angulation seen during late prenatal and early postnatal human development is distinct from that observed for macaques during equivalent periods (see Sirianni and Van Ness, 1978; Sirianni and Newell-Morris, 1980). Studies show that the cranial base angle of Macaca nemestrina remains essentially unchanged during most of the fetal period (Sirianni and Newell-Morris, 1980) 334 N. JEFFERY AND F. SPOOR Fig. 5. Bivariate plots of interpetrosal angle (IPA) and cranial base angle (CBA) against index of differential encephalisation (IDE). Two distinct phases are shown for both variables in relation to IDE, and are separated by mean abscissa value (vertical line). Analysis of IPA against IDE for the first phase realized a significant association, for which the reduced major axes regression line fitting is given (diagonal line). Fig. 4. Bivariate plots of (a) cranial base angle (CBA) against relative endocranial size (IRE), and (b) interpetrosal angle (IPA) against IRE. Reduced major axes regression line fittings are shown for each plot. and subsequently retroflexes by some 10° after birth (Sirianni and Van Ness, 1978). In contrast, the human base retroflexes before birth and subsequently flexes postnatally. It is plausible that the process of human fetal retroflexion is a precocious variant of the retroflexion seen in postnatal nonhuman primates. A possible explanation for such precocity may be found in comparisons between the order in which the basicranial synchondroses fuse in modern humans and macaques. The synchondroses are major centers for growth of the midline primate basicranium and major determinants of its architecture (see Scott, 1958; Lieberman and McCarthy, 1999). In macaques, the spheno-ethmoidal synchondrosis (MES) fuses at birth (Michejda and Lamey, 1971). The midsphenoidal synchondrosis (MSS) becomes inactive during the first few months of postnatal life (Melsen, 1971; Giles et al., 1981) and fuses at around 2 years (Michejda, 1971, 1972; Sirianni, 1985). The spheno-occipital synchondrosis (SOS) only normally fuses in adult macaques (Melsen, 1969; Sirianni, 1985). In modern humans, the MSS is the first to fuse and does so late in fetal life (Ford, 1958; Hoyte, 1971, 1975; Kodama, 1976a,b; Shapiro and Robinson, 1980). This is followed by fusion of the SES, which normally remains patent for up to 6 postnatal years (Scott, 1958). Like macaques, the modern human SOS fuses during adulthood (Brash, 1953; Scott, 1958). These observations demonstrate that the temporal sequence of fusion in the macaque moves posteriorly along the basicranium, starting with the SES and ending with the SOS. By contrast, the timing of SES and MSS fusion is reversed during human development. Perhaps it is the earlier fusion of the MSS that pushes retroflexion into the human fetal period, allowing for postnatal flexion up and until SES fusion (Jeffery, 1999). This explanation remains to be evaluated. Interpetrosal angle. The findings of the present study show that the petrous bones gradually reorient coronally by some 15° over the period investigated. Ford (1956) noted that the angle between the long axes of the otic capsules, which surround the fetal inner ear apparatus, is about 90°. This is close to the mean interpetrosal angle computed in the present study for the interval 27–29 weeks. Ford (1956) also described an increase in distance between the medial margins of the internal auditory meatuses in accordance with overall fetal skull growth. However, he added that after ossification of the otic capsules, at around 18 weeks of gestation, the petrous bones are more sagittally aligned than before. This is consistent with the findings of Lee et al. (1996), who recorded that the angle between the otic capsules rapidly decreases during the period FETAL ENCEPHALIZATION AND BASE ARCHITECTURE 22–30 weeks of gestation. The authors showed that this sagittal reorientation slows, almost to a plateau, later in fetal life. This and the latter finding of Ford (1956) are incongruous with that reported in the present study, which showed a coronal reorientation of the petrous bones beyond 18 weeks and during the period 22–29 weeks. It is worth noting, however, that Lee et al. (1996) measured their angle between line-segments that pass through the otic capsules and an intercept that is fixed at the center of the pituitary fossa in the ventral radiographic view. By forcing an intercept at this point, the authors assumed that its position remains stable relative to the axes of the otic capsules. It is plausible that shifts in the relative position of the pituitary fossa (e.g., Latham, 1972) draw the line segments together and reduces the angle defined by Lee et al. (1996) without regard to petrous orientation. This would appear to be the case. Bossy and Gaillard de Collogny (1965) showed that without a fixed intercept, the angle between the long axes of the otic capsules, each defined by a line segment passing through the cochlea and center of the anterior semicircular canal, increases by some 30°, or 15° relative to the midline, from 10 weeks until birth. This demonstrates coronal as opposed to sagittal reorientation of the otic capsules, and as such is in agreement with the findings of this study. Cranial base lengths. The results of the present study confirm a previous observation that increases in length of the anterior cranial base exceed those of the posterior cranial base by as much as twofold (e.g., Mestre, 1959; Burdi, 1965, 1969; Levihn, 1967; Houpt, 1970; Johnston, 1974; Anagnostopoulou et al., 1988; Eriksen et al., 1995). A similar difference is also seen in macaque fetuses (Sirianni and Newell-Morris, 1980). According to Ford (1956, 1958), the human fetal anterior cranial base has a faster rate of elongation because its direction of midline extension closely matches the anteroposterior vector of brain enlargement, whereas the midline posterior cranial base is inclined to the brain vector and has a slower, skeletal rate of extension. The extension of the midline anterior cranial base in macaques is said to correspond with growth of the face rather than the brain (Sirianni and Swindler, 1979). Two factors that may also underlie differences in the rate of extension between the anterior and posterior parts of the primate cranial base are the positions of growth loci and ossification. The midline anterior cranial base has three major sites for anteroposterior extension, the fronto-ethmoidal suture, sphenoethmoidal synchondrosis, and midsphenoidal synchondrosis. The midline posterior cranial base, on the other hand, only has two, the midsphenoidal and spheno-occipital synchondroses (see Lieberman et al., 2000). Furthermore, the greater part of the posterior cranial base, the basioccipital, ossifies earlier than parts of the anterior cranial base. This reduces the amount of interstitial cartilage growth, which 335 Johnston (1974) suggested is more prodigious than that of bone. Brain volumes and differential encephalization. The scale of endocranial, supratentorial, and infratentorial enlargement reported in the present study is consistent with the findings of Jenkins (1921) and Koop et al. (1986). However, note that technical differences between these studies and the present investigation mean that detailed comparisons of absolute volumes are of little value (see Jeffery, 1999). Notwithstanding this, the findings of Jenkins (1921) and Koop et al. (1986) also demonstrate that the rate of supratentorial expansion exceeds that of the infratentorial brain. This is commensurate with the process of differential encephalization noted in this study. Further corroboration comes from studies showing that regions of the fetal brain anterior to the sella have a greater rate of enlargement than regions posterior to the sella (Blechschmidt, 1977; Friede, 1981). More in particular, Guihard-Costa and Larroche (1990) detailed a trend of decreasing infratentorial mass relative to supratentorial mass with age. This trend has an early period of rapid change followed by a gradual decrease of change, almost to a plateau, and is thus similar to the temporal trajectory of differential encephalization noted in the present study. GuihardCosta and Larroche (1990) also showed that the rate of change begins to increase again later in fetal life, as the infratentorial mass increases relative to the supratentorial mass. Adult modern human values of IDE, computed with volumes from Stephan et al. (1981) and Filipek et al. (1989), range from 0.130 – 0.137. These values are larger than the values reported in this study for late fetal life. This confirms the suggestion that differential infratentorial, as opposed to supratentorial, enlargement occurs at some point during the intervening period of the late third fetal trimester to adulthood, since infratentorial volume must increase relative to supratentorial volume for there to be an increase of IDE (infra/supra volume). Indeed, studies show that the cerebellum, which makes up most of the infratentorial volume, starts to develop later than other parts of the fetal brain but finishes its growth earlier, by way of a late fetal to early postnatal period of rapid expansion (Noback and Moss, 1956; Rakic and Sidman, 1970; Dobbing and Sands, 1973). Relative infratentorial enlargement at this time may be augmented by a contemporaneous deceleration of supratentorial growth (Guihard-Costa and Larroche, 1990), which perhaps evolved in modern humans as a safeguard against difficult births of large-headed neonates (see Leutenegger, 1987). Results from the present study showed temporally related increases of relative endocranial size, but relative infratentorial size was not temporally correlated. There are no previous developmental studies with which to compare these findings. However, it is interesting to note that the fetal range of relative endocranial size closely 336 N. JEFFERY AND F. SPOOR matches the range of relative brain sizes observed across extant primates at the hominoid taxonomic level (Spoor, 1997; McCarthy, 2001). Assuming that interspecific analyses provide insights into changes with phylogenetic time, this would suggest that the scale of relative brain enlargement seen during fetal development is similar to that which occurred during human evolution. As a corollary point, it is also worth noting that McCarthy (2001) detailed an adult modern human value for relative brain size of 1.22. This closely matches the value of relative endocranial size observed here in the last few weeks of the period investigated. It may be that the fetal value of relative size is retained into adulthood, or that temporally distinct and possibly different processes afford this likeness, as appears the case with regard to cranial base angle. Hypotheses General spatial-packing. This hypothesis suggests that increases of brain size (volume) relative to length of the midline cranial base create a spatialpacking problem that drives cranial base flexion and coronal reorientation of the petrous bones. The findings from this study appear to support one aspect of the hypothesis while refuting the other. The result for cranial base angle is the antithesis of the predicted outcome. The fetal cranial base retroflexed rather than flexed in relation to increases of relative brain size. This raised the suspicion that the individual trends of these variables with age, nee head size, were superimposed in the computation of the simple rank correlation, which showed an association between relative encephalization and retroflexion. Indeed, the partial correlation, computed with biparietal diameter held constant, suggested that cranial base angle is not directly associated with relative endocranial size. This suggests that relative increases of endocranial size were accommodated elsewhere in the fetal skull, perhaps by enlargement of the calvaria, to allow for a shallower posterior cranial fossa. On first impression, the findings of this study appear to corroborate the idea that spatial packing is the driving force behind petrous reorientation. However, the correlation between relative endocranial size and interpetrosal angle is small. Moreover, both these measures also covaried with age. Given the compounding of variations seen in relation to age in the case of cranial base angle, it also seemed pertinent to question the robusticity of the rank correlation here. This misgiving is justified. The partial correlation shows that interpetrosal angle is not associated with relative endocranial size while controlling for biparietal diameter. Thus, the partial correlation coefficients indicate that cranial base angulation and petrous reorientation were correlated to increases in vault size rather than directly to relative endocranial size. As such, the general spatial-packing hypothesis is not corroborated by the findings of this study. Infratentorial spatial-packing. This hypothesis predicts that increases of infratentorial volume, i.e., cerebellum and brain-stem volume, relative to the length of the posterior cranial base create a spatial-packing problem within the posterior cranial fossa that is alleviated by cranial base flexion and coronal reorientation of the petrous bones. This study corroborated that increases in relative infratentorial size are associated with slower growth of the posterior cranial base (Moss, 1958; Dean, 1988), but these were not correlated to changes in cranial base angle and petrous orientation. Thus, the infratentorial spatial-packing hypothesis is not corroborated for the period studied. This is not to say that the hypothesis will be invalidated for all of ontogeny. As previously noted, the human cerebellum undergoes a relative increase in rate of enlargement during a later period of development than that investigated here. This latent and more pronounced cerebellar expansion may create a spatial-packing problem that affects a change of basicranial morphology, as per the hypothesis. More work is needed in this area to evaluate the temporal scope of mechanisms for change. Differential encephalization. Marked changes of proportions of the brain were observed in the present study. Supratentorial volume was shown to increase in relation to infratentorial volume. The hypothesis predicts that cranial base angle decreases and interpetrosal angle increases with the process of differential encephalization. However, testing this concept proved complex because of the biphasic trajectories of the plots. This problem was circumvented by subdividing the analyses according to the phases observed. While not objective, this approach is straightforward, and more importantly is effective in describing the individual phases shown. Subsequent analysis of individual phases showed that changes of cranial base angle are not associated with differential encephalization during the period investigated. It was also shown that changes of interpetrosal angle are not associated with differential encephalization in early fetal life (i.e., 10 –20 weeks). These findings do not corroborate the hypothesis. However, the hypothesis was seemingly corroborated by the small correlation observed between changes of interpetrosal angle and differential encephalization later in fetal life (i.e., 20 –29 weeks). Again, such a small correlation and the covariation of these measures with age raised the suspicion that temporal colinearities were compounded in the analysis. This was confirmed by the partial correlations while holding for biparietal diameter. Thus, the findings are not consistent with the hypothesis that differential encephalization drives cranial base flexion or petrous reorientation during human fetal life, at least during the period investigated. FETAL ENCEPHALIZATION AND BASE ARCHITECTURE Interpretations The major finding of this study is that changes of cranial base angle and interpetrosal angle are independent of increases of absolute and relative endocranial sizes. Even if we disregard the partial correlations, there is little support for any of the hypotheses put forward at the beginning of this study. Rather, the evidence suggests that the observed changes arise from temporally related factors that at present remain undetermined. The obvious question is: what are these factors? Before considering answers to this question, we must point out the major limitations of this study. Results from this study, as well as those from other prenatal studies, only cover part of the gestational period and an even smaller part of the overall period of human development. Mechanistic hypotheses remain to be tested over time spans such as the late fetal or perinatal periods. This caveat, combined with the finding that many of the trends shown here are polyphasic (e.g., cranial base angulation and differential encephalization), raises the possibility that the hypotheses may prove valid explanations for one period of development but not for another. In other words, the findings and conclusions of this study can only be related with any certainty to human fetal material from the same period. Any extrapolation from this is useful in guiding further study and formulating ideas, both developmental and evolutionary, but such interpretations remain tenuous until a study of wider scope can be undertaken. In addition, the aims of the present study were strictly focused on the effects of brain enlargement, in its various forms. Other potential mechanistic processes were not considered. For example, the size of the pharynx has been linked with changes of cranial base angle during human development and with the degree of flexion seen in adult modern humans compared with that of other primates (see Ross and Henneberg, 1995; Lieberman et al., 2001). Such mechanisms may work against or in tandem with the effects of brain enlargement, or even in opposing phases over ontogenetic time. There is still some way to go before all possible mechanistic explanations for changes of the human fetal cranial base, and their permutations, are exhausted. The fetal basicranium appears to remain unperturbed by considerable changes in both brain size and shape, suggesting that the pattern of variation taken by the fetal cranial base is partially a consequence of variables other than structural demands. Notwithstanding those mechanisms not yet tested, the demonstration of independence implies that fetal basicranial morphology is regulated at some level by internal, i.e., intrinsic, factors in preference to external influences. It is unclear from this study whether “intrinsic” regulation corresponds to direct gene control or the knock-on effects of gene control separated from the initial expression by intervening levels of increasing biological complexity (i.e., from 337 gene expression, through signalling, cellular activity, and tissue morphogenesis, to architectural variations). For example, it is plausible that base flexion and petrous reorientation are directly regulated by genes that represent a modern human adaptation, selected for over subsequent ancestral ontogenies in response to a mechanical force like brain enlargement but which are now exapted for some other purpose and are largely independent of brain size. Perhaps more plausible is that there is no direct association, and that genetic control is mediated through processes like the position and activity of growth loci within the cranial base. That way, similar genotypes are gradually overlaid with the structural demands of growth during ontogeny to produce different phenotypes. Some insight can be gained by exploring the disparity between results from the interspecific studies cited and the ontogenetic findings reported here. In nearly every case, the findings of this study contradict those obtained from interspecific analyses of extant primates (e.g., Ross and Ravosa, 1993; Spoor, 1997; McCarthy, 2001). This apparent dichotomy may be an artifact of the comparative method used in these interspecific studies. Like the compounding of variables by their relationship with age seen in this study, interspecific trends can be compounded with the phylogenetic relationships among the often closely related primate species investigated (see Harvey and Pagel, 2000). Indeed, Lieberman et al. (2000) noted that with degrees of freedom adjusted for phylogeny across extant anthropoids, the correlation between differential encephalization (telencephalon/brain-stem) and cranial base angle is insignificant. The authors indicated that the unadjusted correlation was significant. Their modified result is consistent with the finding reported here, showing that cranial base angle is not significantly correlated with differential encephalization (infra/ supratentorial). However, this does not resolve all differences. Many interspecific results remain at odds with those reported here, even after adjustment. For example, Lieberman et al. (2000) also showed that while the adjusted correlation between relative brain size (IRE) and cranial base angle is less significant than that reported by Ross and Ravosa (1993), it is still robust enough to support the spatial-packing hypothesis. Such seemingly irreconcilable differences point to a more fundamental incongruity, whereby ontogenetic and interspecific studies produce results pertaining to different processes (e.g., Zuckerman, 1955; Sirianni and Swindler, 1979). Ontogeny and phylogeny are clearly interconnected, but to what extent and is this link constant? Clearly, the interchange occurs at the adult intraspecific level, where variations during ontogeny provide a suite of variations in the sexually mature which, if selected for and heritable, can evolve into new morphologies (see Gould and Lewontin, 1979). This implies that developmental findings differ in 338 N. JEFFERY AND F. SPOOR that they pertain to a dynamic process of reversal between phylogeny and the changes responsible for intraspecific variation. To put it another way, ontogeny is the stochastic accumulation of somatic variations against the backdrop of decreasing inherited control of morphology. The core of this idea is wellestablished in the literature (e.g., Baker, 1941; Scott, 1954; Manson, 1968; Dorenbos, 1972; Tuckett and Morriss-Kay, 1985). According to this concept, the fetal basicranium is subject to more intrinsic control and less mechanistic control than the postnatal basicranium. Thus, the period of ontogeny examined determines the findings of the study with regard to the testing of mechanistic hypotheses. This can explain the adult interspecific/human fetal inconsistencies at two levels. First, it can be proposed that inconsistency results from generally tighter regulation of primate fetal development compared with the adult primate form. In other words, the adult interspecific studies, by comparison, highlight biomechanistic-driven changes across taxa, while the results presented here indicate that intrinsic controls are not yet sufficiently relaxed in the early fetus to allow for such marked relationships. This predicts that intraspecific studies of fetal development particularly at the earliest stages in other primate taxa, will yield similar inconsistencies with the adult interspecific evidence. Second, it can also be proposed that within this general scheme of primate fetal development, different temporal patterns in the relaxation of control exist between species, thereby contributing to interspecific differences later in life, and that modern humans exhibit a comparatively extended period of tighter morphological control. This predicts that developmental studies across primate taxa will demonstrate that certain species group trends are consistent with adult interspecific trends owing to the comparatively earlier relaxation of control, perhaps even at the fetal stage of life. Despite such insights, a number of questions remain unanswered. In particular, the essential problem remains of distinguishing adaptations incorporated into the genome by successive ontogenies from those morphologies that do not require assimilation into the genome, because the ubiquitous nature of mechanical laws can be relied on to deliver a sufficient phenotype, and from morphologies produced by the mechanistic knock-on effects of adaptive change. Only by combining ontogenetic and interspecific analyses, tempered with evidence from the fossil record, can we hope to obtain a clearer picture of the processes underlying human evolution, and in particular the factors underlying the derived condition of the modern human skull. CONCLUSIONS Data were collected from a sample of human fetuses ranging from 10 –29 weeks of gestation in order to test hypotheses that cranial base angle and petrous orientation follow from increases of relative endocranial size, increases of relative infratentorial size, or differential encephalization. We found that: 1) The cranial base retroflexes during the human fetal period investigated. Considering findings from previous studies, it was determined that the process of basicranial angulation is polyphasic, and that fetal values of cranial base angle are not retained into adulthood. 2) It was also revealed that the petrous bones coronally reorientate, relative brain size increases, and supratentorial volume differentially enlarges over the period investigated. 3) Results show distinct growth rates on either side on the tentorium cerebelli for both cranial base lengths and endocranial volumes. Those components superoanterior to the tentorium cerebelli increase at rates exceeding those of inferoposterior components. 4) Lastly and most importantly, the findings of this study show that cranial base angulation and petrous reorientation are associated with increases of vault size, which represent age, rather than increases of relative brain sizes. Given the notable increases of relative sizes observed, this suggests that fetal basicranial morphology is regulated more intrinsically than proposed by the hypotheses tested, though many other mechanistic explanations remain to be discounted. ACKNOWLEDGMENTS We thank B. Berkovitz, W. Birch, M. Bird, P. Kinchesh, and P. O’Higgins for permission and help with imaging the material. We also thank four anonymous referees for their comments and suggestions. LITERATURE CITED Anagnostopoulou S, Karamaliki DD, Spyropoulos MN. 1988. Observations on the growth and orientation of the anterior cranial base in the human foetus. Eur J Orthod 10:143–148. Ashton EH. 1957. Age changes in the basicranial axis of the anthropoidea. Proc Zool Soc Lond 129:61–74. Baker LW. 1941. The influence of the formative dental organs on the growth of bones and the face. Am J Orthod 27:489 –506. Biegert J. 1957. Der Formwandel des Primatenschädels und seine Beziehungen zur ontogenetischen Entwicklung und den phylogenetischen Spezialisationen der Kopforgane. Gegenbauers Morphol Jahrb 98:77–199. Biegert J. 1963. The evaluation of characteristics of the skull, hands, and feet for primate taxonomy. In: Washburn SL, editor. Classification and human evolution. London: Methuen. p 116 – 145. Blechschmidt E. 1977. The beginnings of human life. Translated by Transemantics, Inc. New York: Spinger-Verlag. Bossy J, Gaillard de Collogny L. 1965. Orientation comparée de la cochlée et du canal semi-circulaire anterieur chez le foetus. J Fr Otorhinolaryngol 14:727–735. Brash JC. 1953. Cunningham’s text-book of anatomy. London: Oxford University Press. Burdi AR. 1965. Sagittal growth of the nasomaxillary complex during the second trimester of human prenatal development. J Dent Res 44:112–125. Burdi AR. 1969. Cephalometric growth analyses of the human upper face region during the last two trimesters of gestation. Am J Anat 125:113–122. FETAL ENCEPHALIZATION AND BASE ARCHITECTURE Cameron J. 1924. The cranio-facial axis of Huxley. Part I. Embryological considerations. Trans R Soc Can 18:115–123. Cameron J. 1930. The human and comparative anatomy of Cameron’s craniofacial axis. J Anat 64:324 –336. Chitty LS, Altman DG, Henderson A, Campbell S. 1994. Charts of fetal size: 2. Head measurements. Br J Obstet Gynaecol 101: 35– 43. Cousin RP. 1969. Étude en projection sagittale de crânes d’enfants orientés dans les axes vestibulaires. Ph.D. dissertation, University of Paris. Dabelow A. 1931. Über Korrelationenin der phylogenetischen Entwicklung der Schädelform II. Die Beziehungen zwischen Gehirn und Schädelbasisform bei den Mammalien. Gegenbauers Morphol Jahrb 67:84 –133. Dean MC. 1988. Growth processes in the cranial base of hominoids and their bearing on morphological similarities that exist in the cranial base of Homo and Paranthropus. In: Grine FE, editor. Evolutionary history of the “robust” australopithecines. New York: Aldine de Gruyter. p 107–112. Dean MC, Wood BA. 1981. Metrical analysis of the basicranium of extant hominoids and Australopithecus. Am J Phys Anthropol 54:63–71. Dean MC, Wood BA. 1984. Phylogeny, neoteny and growth of the cranial base in Hominoids. Folia Primatol (Basel) 43:157–180. De Beer GR. 1937. The development of the vertebrate skull. Oxford: Clarendon Press. Diewert VM. 1983. A morphometric analysis of craniofacial growth and changes in spatial relations during secondary palatal development in human embryos and fetuses. Am J Anat 167:495–522. Diewert VM. 1985. Development of human craniofacial morphology during the late embryonic and early fetal periods. Am J Orthod 88:64 –76. Dimitriadis AS, Haritanti-Kouridou A, Antoniadis K, Ekononmou L. 1995. The human skull base angle during the second trimester of gestation. Neuroradiology 37:68 –71. Dobbing J, Sands J. 1973. Quantitative growth and development of human brain. Arch Dis Child 48:757–767. Dorenbos J. 1972. In vivo cerebral implantation of the anterior and posterior halves of the spheno-occipital synchondrosis in rats. Arch Oral Biol 17:1067–1072. Du Brul EL. 1977. Early hominid feeding mechanisms. Am J Phys Anthropol 47:305–320. Duckworth WLH. 1915. Morphology and anthropology. Cambridge: Cambridge University Press. Enlow DH. 1976. The prenatal and postnatal growth of the human basicranium. In: Bosma JF, editor. Symposium on development of the basicranium. Bethesda: US Government DHEW. Publication no. NIH 76-989. p 192–205. Enlow DH, Hunter WS. 1968. The growth of the face in relation to the cranial base. Rep Congr Eur Orthod Soc 44:321–335. Enlow DH, Moyers RE. 1971. Growth and architecture of the face. J Am Dent Assoc 82:763–774. Erdoglija LJ. 1989. Dynamics of the cranial base angle changes during the second trimester of the normal intrauterine growth and development. Bilt Udruz Ortod Jugosl 22:7–14. Erdoglija LJ. 1990. Dynamics of changes of anteroposterior jaw positions relative to the cranial base during the second trimester of normal intrauterine growth. Bilt Udruz Ortod Jugosl 23:59 – 68. Eriksen E, Bach-Petersen S, van den Eynde B, Solow B, Kjaer I. 1995. Midsagittal dimensions of the prenatal human cranium. J Craniofac Genet Dev Biol 15:44 –50. Filipek PA, Kennedy DN, Caviness VS Jr, Rossnick SL, Spraggins TA, Starewicz PM. 1989. Magnetic resonance imaging based brain morphometry: development and application to normal subjects. Ann Neurol 25:61– 67. Ford EHR. 1956. The growth of the foetal skull. J Anat 90:63–72. Ford EHR. 1958. Growth of the human cranial base. Am J Orthod 44:498 –506. Friede H. 1981. Normal development and growth of the human neurocranium and cranial base. Scand J Plast Reconstr Surg 15:163–169. 339 Garstang W. 1922. The theory of recapitulation: a critical restatement of the biogenetic law. Zool J Linn 35:81–101. Giles WB, Philips CL, Joondeph DR. 1981. Growth in the basicranial synchondroses of the adolescent Macaca mulatta. Anat Rec 199:259 –266. Gould SJ. 1977. Ontogeny and phylogeny. London: Harvard University Press. Gould SJ, Lewontin RC. 1979. The spandrels of San Marco and the Panglossian paradigm: a critque of the adaptationist programme. Proc R Soc Lond [Biol] 205:581–598. Guihard-Costa AM, Larroche JC. 1990. Differential growth between the fetal brain and its infratentorial part. Early Hum Dev 23:27– 40. Guihard-Costa AM, Larroche JC. 1992. Growth velocity of some fetal parameters. I. Brain weight and brain dimensions. Biol Neonate 62:309 –316. Harvey P, Pagel M. 2000. The comparative method in evolutionary biology. Oxford: Oxford University Press. Herring SW. 1993. Epigenetic and functional influences on skull growth. In: Hanken J, Hall BK, editors. The skull: development. Chicago: Chicago University Press. p 153–207. Hofer HO. 1969. On the evolution of the craniocerebral topography in primates. Ann NY Acad Sci 162:15–24. Hoyte DA. 1971. Mechanisms of growth in the cranial vault and base. J Dent Res 50:1447–1461. Hoyte DA. 1975. A critical analysis of the growth in length of the cranial base. Birth Defects 11:255–282. Houpt MI. 1970. Growth of the craniofacial complex of the human fetus. Am J Orthod 58:373–383. Jeffery N. 1999. Fetal development and evolution of the human cranial base. Ph.D. dissertation, University College, London. Jenkins GB. 1921. Relative weight and volume of the component parts of the brain of the human embryo at different stages of development. Contrib Embryol 29:41– 60. Jerison HJ. 1982. Allometry, brain size and convolutedness. In: Armstrong E, Falk D, editors. Primate brain evolution: methods and concepts. New York: Plenum Press. p 77– 84. Johnston LE. 1974. A cephalometric investigation of the sagittal growth of the second trimester fetal face. Anat Rec 178:623– 630. Keith A. 1910. Description of a new craniometer and of certain age changes in the anthropoid skull. J Anat 44:251–270. Knowles CC. 1963. The influence of cranial base structure on the orientation of the middle third of the face. Dent Pract Dent Rec 13:531–542. Kodama G. 1976a. Developmental studies on the presphenoid of the human sphenoid bone. In: Bosma JF, editor. Symposium on development of the basicranium. Bethesda: US Government DHEW. Publication no. NIH 76-989. p 141–155. Kodama G. 1976b. Developmental studies on the body of the human sphenoid bone. In: Bosma JF, editor. Symposium on development of the basicranium. Bethesda: US Government DHEW. Publication no. NIH 76-989. p 156 –165. Koop M, Rilling G, Herrmann A, Kretschmann HJ. 1986. Volumetric development of the fetal telencephalon, cerebral cortex, diencephalon, and rhombencephalon including the cerebellum in man. Bibl Anat 28:53–78. Kvinnsland S. 1971. The sagittal growth of the foetal cranial base. Acta Odontol Scand 29:699 –715. Latham RA. 1972. The sella point and postnatal growth of the human cranial base. Am J Orthod 61:156 –162. Lee SK, Kim YS, Jo YA, Seo JW, Chi JG. 1996. Prenatal development of cranial base in normal Korean fetuses. Anat Rec 246:524 –534. Leutenegger W. 1987. Neonatal brain size and neurocranial dimensions in Pliocene hominids: implications for obstetrics. J Hum Evol 16:291–296. Levihn WC. 1967. A cephalometric roentgenographic cross-sectional study of the craniofacial complex in fetuses from 12 weeks to birth. Am J Orthod 53:822– 848. Lieberman DE, McCarthy RC. 1999. The ontogeny of cranial base angulation in humans and chimpanzees and its implications for reconstructing pharyngeal dimensions. J Hum Evol 36:487– 517. 340 N. JEFFERY AND F. SPOOR Lieberman DE, Ross CF, Ravosa MJ. 2000. The primate cranial base: ontogeny, function, and integration. Am J Phys Anthropol Yrbk 43:117–169. Lieberman DE, McCarthy RC, Hiiemae KM, Palmer JB. 2001. Ontogeny of postnatal hyoid and larynx descent in humans. Arch Oral Biol 46:117–128. Mandarim-de-Lacerda CA, Alves MU. 1992. Growth of the cranial bones in human fetuses (2nd and 3rd trimesters). Surg Radiol Anat 14:125–129. Manson JD. 1968. A comparative study of the postnatal growth of the mandible. London: Henry Kempton. McCarthy RC. 2001. Anthropoid cranial base architecture and scaling relationships. J Hum Evol 40:41– 66. Melsen B. 1969. Time of closure of the spheno-occipital synchondrosis determined on dry skulls. A radiographic craniometric study. Acta Odontol Scand 27:73–90. Melsen B. 1971. The postnatal growth of the cranial base in Macaca rhesus analyzed by the implant method. Tandlaegebladet 75:1320 –1329. Mestre JC. 1959. A cephalometric appraisal of cranial and facial relationships at various stages of human fetal development. Am J Orthod 45:473. Michejda M. 1971. Ontogenetic changes of the cranial base in Macaca mulatta. Proceedings of the Third International Conference on Primatology, Zurich, 1970. Volume 1. p 215–225. Michejda M. 1972. The role of the basicranial synchondroses in flexure processes and ontogenetic development of the skull base. Am J Phys Anthropol 37:143–150. Michejda M, Lamey D. 1971. Flexion and metric age changes of the cranial base in the Macaca mulatta. I. Infants and juveniles. Folia Primatol (Basel) 14:84 –94. Moss ML. 1958. The pathogenesis of artificial cranial deformation. Am J Phys Anthropol 16:269 –285. Moss ML, Noback CR, Robertson GG. 1956. Growth of certain human fetal cranial bones. Am J Anat 98:191–204. Müller F, O’Rahilly R. 1980. The human chondrocranium at the end of the embryonic period, proper, with particular reference to the nervous system. Am J Anat 159:33–58. Noback CR, Moss ML. 1956. Differential growth of the human brain. J Comp Neurol 105:539 –555. Plavcan JM, German RZ. 1995. Quantitative evaluation of craniofacial growth in the third trimester human. Cleft Palate Craniofac J 32:394 – 404. Putz VR. 1974. Skull configuration and pyramids. On the position of pyramids within the cranial base. Anat Anz 135:252–266. Rakic P, Sidman RL. 1970. Histogenesis of cortical layers in human cerebellum, particularly the lamina dissecans. J Comp Neurol 139:473–500. Ross CF, Henneberg M. 1995. Basicranial flexion, relative brain size, and facial kyphosis in Homo sapiens and some fossil hominids. Am J Phys Anthropol 98:575–593. Ross CF, Ravosa MJ. 1993. Basicranial flexion, relative brain size, and facial kyphosis in nonhuman primates. Am J Phys Anthropol 91:305–324. Scheuerle AE. 1995. Recent advances in craniofacial genetics. J Craniofac Surg 6:440 – 442. Schultz AH. 1955. The position of the occipital condyles and of the face relative to the skull base in primates. Am J Phys Anthropol 13:97–120. Scott JH. 1954. The growth of the human face. Proc R Soc Med 47:9147–9158. Scott JH. 1958. The cranial base. Am J Phys Anthropol 16:319 – 348. Shapiro R, Robinson F. 1980. The embryogenesis of the human skull. London: Harvard University Press. Sherwood RJ, Meindl RS, Robinson HB, May RL. 2000. Fetal age: methods of estimation and effects of pathology. Am J Phys Anthropol 113:305–315. Sirianni JE. 1985. Nonhuman primates as models for human craniofacial growth. In: Alan R, editor. Nonhuman primate models for human growth and development. New York: Liss, Inc. p 95–124. Sirianni JE, Newell-Morris L. 1980. Craniofacial growth of fetal Macaca nemestrina: a cephalometric roentgenographic study. Am J Phys Anthropol 53:407– 421. Sirianni JE, Swindler DR. 1979. A review of postnatal craniofacial growth in old world monkeys and apes. Am J Phys Anthropol Yrbk 22:80 –104. Sirianni JE, Van Ness AL. 1978. Postnatal growth of the cranial base in Macaca nemestrina. Am J Phys Anthropol 49:329 –340. Sokal RR, Rohlf FJ. 1995. Biometry. New York: W.H. Freeman and Co. p 593– 609. Sperber GH. 1992. Current concepts in embryonic craniofacial development. Crit Rev Oral Biol Med 4:67–72. Sperber GH. 2001. Craniofacial development. London: B.C. Decker, Inc. p 96 –97. Spoor F. 1997. Basicranial architecture and relative brain size of Sts 5. Australopithecus africanus and other Plio-Pleistocene hominids. S Afr J Sci 93:182–187. Spoor F, Zonneveld FW. 1995. Morphometry of the primate bony labyrinth: a new method based on high-resolution computed tomography. J Anat 186:271–286. Spoor F, Zonneveld FW. 1998. Comparative review of the human bony labyrinth. Am J Phys Anthropol 27:211–251. Spoor F, Jeffery N, Zonneveld FW. 2000a. Imaging skeletal growth and evolution. In: O’Higgins P, Cohen M, editors. Development, growth and evolution: implications for the study of hominid skeletal evolution. London: Academic Press. p 123– 161. Spoor F, Jeffery N, Zonneveld FW. 2000b. Using diagnostic radiology in human evolutionary studies. J Anat 197:61–76. Stephan H, Frahm H, Baron G. 1981. New and revised data on volumes of brain structures in insectivores and primates. Folia Primatol (Basel) 35:1–29. Strait DS. 1999. The scaling of basicranial flexion and length. J Hum Evol 37:701–719. Strait DS, Ross CF. 1999. Kinematic data on primate head and neck posture: implications for the evolution of basicranial flexion and an evaluation of registration planes used in paleoanthropology. Am J Phys Anthropol 108:205–222. Thesleff I. 1998. The genetic basis of normal and abnormal craniofacial development. Acta Odontol Scand 56:321–325. Thorogood P. 1987. Mechanisms of morphogenetic specification skull development. In: Wolff JR, Sievers J, Berry M, editors. Mesenchymal-epithelial interactions in neural development. Berlin: Spinger-Verlag. p 141–152. Thorogood P. 1988. The developmental specification of the vertebrate skull. Development 103:141–153. Thorogood P. 1993. Differentiation and morphogenesis of cranial skeletal tissues. In: Hanken J, Hall BK, editors. The skull: development. Chicago: Chicago University Press. p 112–153. Tuckett F, Morriss-Kay GM. 1985. The ontogenesis of cranial neuromeres in the rat embryo. II. A transmission electron microscope study. J Embryol Exp Morphol 88:231–247. van den Eynde B, Kjaer I, Solow B, Graem N, Kjaer TW, Mathiesen M. 1992. Cranial base angulation and prognathism related to cranial and general skeletal maturation in human fetuses. J Craniofac Genet Dev Biol 12:22–32. Virchow R. 1857. Untersuchungen über die Entwicklung des Schädelgrundes im gesunden und frankhaften Zustande. Berlin: Gesichissiblung und Gehirnbru. Zuckerman S. 1955. Age changes in the basicranial axis of the human skull. Am J Phys Anthropol 13:521–539.