View Full Version : Basicranial influence on overall cranial shape

Saturday, October 23rd, 2004, 03:36 PM
Daniel E. Lieberman - Department of Anthropology,
The George Washington University, 2110 G St, NW, Washington, DC 20052,

This study examines the extent to which the major dimensions of the
cranial base (maximum length, maximum breadth, and flexion)
interact with brain volume to influence major proportions of the
neurocranium and face. A model is presented for developmental
interactions that occur during ontogeny between the brain and the
cranial base and neurocranium, and between the neurobasicranial
complex (NBC) and the face. The model is tested using exocranial
and radiographic measurements of adult crania sampled from five
geographically and craniometrically diverse populations. The results
indicate that while variations in the breadth, length and flexion of the
cranial base are mutually independent, only the maximum breadth of
the cranial base (POB) has significant effects on overall cranial
proportions, largely through its interactions with brain volume which
influence NBC breadth. These interactions also have a slight influence
on facial shape because NBC width constrains facial width, and
because narrow-faced individuals tend to have antero-posteriorly
longer faces relative to facial breadth than wide-faced individuals.
Finally, the model highlights how integration between the cranial
base and the brain may help to account for the developmental basis of
some morphological variations such as occipital bunning. Among
modern humans, the degree of posterior projection of the occipital
bone appears to be a consequence of having a large brain on a
relatively narrow cranial base. Occipital buns in Neanderthals, who
have wide cranial bases relative to endocranial volume, may not be
entirely homologous with the morphology occasionally evident in
Homo sapiens.
 2000 Academic Press
Journal of Human Evolution (2000) 38, 291–315
doi: 10.1006/jhev.1999.0335
Available online at http://www.idealibrary.com on
It has long been known that the cranial base,
vault and face derive from embryologically
distinct regions (the basicranium, neurocranium
and splanchnocranium) but that these
regions grow in a morphologically integrated
manner through numerous developmental
and functional interactions (de Beer, 1937;
Moss & Young, 1960; Enlow, 1968, 1990;
Cheverud, 1982; Sperber, 1989). Although
these interactions occur as the result of
many morphogenetic (e.g., neural) and
functional (e.g., masticatory, respiratory)
stimuli, the role of the cranial base in influencing
overall cranial shape merits special
consideration. Developmentally, the basicranium
differs from the neurocranium and
splanchnocranium in several important
respects. Unlike the rest of the skull, which
develops intramembranously from neural
crest-derived tissue, the basicranium mostly
grows from endochondral ossification processes
in which mesodermally-derived cartilaginous
precursors (the chondrocranium)
develop in utero and are gradually replaced
by bone after birth (Sperber, 1989). The
basicranium is also the first region of the
0047–2484/00/020291+25$35.00/0  2000 Academic Press
skull to reach adult size (Moore & Lavelle,
1974), and it is the structural foundation of
many aspects of craniofacial architecture.
The cranial base forms the platform upon
which the rest of the skull grows and
attaches (see Biegert, 1963), and it provides
and protects the crucial foramina through
which the brain connects to the face and the
rest of the body. These aspects of cranial
base growth and function may account for
its apparent morphological and developmental
conservatism in mammals compared
to other regions of the skull (de Beer, 1937;
Bosma, 1976 and references therein;
Sperber, 1989: 117). Consequently, a
number of recent phylogenetic studies of
hominids (e.g., Olson, 1981; Lieberman,
1995; Lieberman et al., 1996; Strait et al.,
1997; Strait, 1998) have proposed that variations
in cranial base morphology may be
better indicators of taxonomy and phylogeny
than neurocranial or facial characters.
Some studies (e.g., Skelton &
McHenry, 1992; Lieberman et al., 1996;
Strait et al., 1997) have specifically examined
cladograms that emphasize the importance
of basicranial traits by grouping
neurocranial and facial characters into functional
complexes. However, it must be
stressed that whether or not the basicranium
is a better source of phylogenetic data remains
open to question. Basicranial, neurocranial,
and facial dimension show similar
levels of heritability within the primate skull
(e.g., Sjøvold, 1984; Cheverud & Buikstra,
1982; Cheverud, 1995) and appear to be
equally well (or poorly) integrated with
other dimensions or features in primate
phylogeny (Strait, 1998).
This study tests the extent to which variations
in the major dimensions of the cranial
base (length, breadth and flexion) may influence
several aspects of the shape of the face
and cranial vault in humans. After presenting
a model for interactions between the
basicranium, neurocranium and face during
growth, we test two sets of hypotheses. First,
it is predicted that variations in the shape of
the human neurocranium are influenced by
interactions between two factors: (1) variations
in the shape of the basicranium upon
which the neurocranium grows, and (2)
endocranial expansion driven by brain
growth. Because the face is displaced in a
forward and downward trajectory from the
basicranium and neurocranium, we also test
the related hypothesis that variations in
basicranial and neurocranial breadth constrain
upper- and mid-facial breadth, and
hence influence other aspects of facial morphology,
especially facial depth and height
(Weidenreich, 1941; Enlow & Bhatt, 1984;
Enlow, 1990). These hypotheses are examined
using a pooled-sex sample of adult
anatomically modern Homo sapiens from five
geographically and craniometrically diverse
populations in order to examine as wide a
range of cranial variation as possible. This
study does not examine inter-population or
intra-population variability. A future goal is
to test the model within populations and
between sexes, but larger samples sizes are
required than presently available (see below).
The results of these analyses are also used
to examine variation in occipital ‘‘bunning’’
among Upper Pleistocene hominids and
recent human populations. Occipital buns,
which have been suggested to be important
for testing phylogenetic hypotheses about
recent human evolution, are defined as
posteriorly-directed projections of the
occipital beyond the nuchal plane that result
in a distinctive swollen morphology when
viewed in norma lateralis (Ducros, 1967;
Trinkaus & LeMay, 1982). Bunning has
been suggested to be a derived Neanderthal
character the presence of which in some
early modern humans from Europe indicates
regional continuity (Smith, 1984; Frayer,
1992a, 1992b; Frayer et al., 1993; Wolpoff,
1996). However, Trinkaus & LeMay (1982)
and Lieberman (1995) have suggested
that bunning may be a developmental
consequence of posteriorly-directed cranial
292 . .  ET AL.
vault expansion that occurs in very largebrained
hominids, such as Neanderthals
or Upper Pleistocene modern humans, in
which a relatively narrow cranial base constrains
lateral vault expansion. If this
hypothesis is correct, then the degree of
occipital projection among adult recent
humans as well as Pleistocene hominids
should be correlated with the ratio of endocranial
volume relative to cranial base
In order to investigate the relationship
between basicranial dimensions and overall
skull shape, it is useful to review several
aspects of craniofacial development, focusing
on Enlow’s (1990) model of the ontogenetic
interactions between the basicranium,
neurocranium, and splanchnocranium
(illustrated in Figure 1). The basicranium is
defined here as the portion of the skull
which derives from the chondrocranium and
which grows through endochondral ossification.
As the basicranium grows, it elongates
and flexes in the spheno-ethmoid,
mid-sphenoid, and spheno-occipital synchondroses
(Scott, 1958). Increases in
basicranial breadth and length also occur in
sutures (e.g., the occipito-mastoid), and the
endocranial fossae of the basicranium
deepen through drift in which the resorption
and deposition occur along the superior and
inferior surfaces, respectively (Enlow,
1990). In contrast, the neurocranium grows
entirely from intramembraneous ossification
processes without any cartilaginous precursors.
Intramembranous osteogenesis of the
neurocranium occurs within the outer portion
of ectomeninx membrane that differentiates
from the dural meninges (Friede,
1981).1 During normal growth in humans,
the upper half of the neurocranium enlarges
mainly from deposition within the cranial
sutures, although some resorption does take
place; its lower half also expands through
drift in which the external (ectocranial) surface
is depository and the internal (endocranial)
surface is resorptive (Duterloo &
Enlow, 1970). Therefore, as Figure 1 illustrates,
the basicranium and neurocranium
grow in tandem in a rapid neural growth
trajectory, forming a highly integrated
morphological unit, the neuro-basicranial
complex (NBC).
This model of cranial growth assumes that
overall shape of the NBC has two primary
influences: the shape of the brain, and the
shape of the basicranium. This integrated
growth occurs through many processes, the
most important of which are sutural expansion,
synchondroseal deposition, drift, and
flexion. As the brain expands, it generates
tension along the endocranial surface of the
neurocranial cavity, activating osteoblast
deposition within intra-sutural periosteum
throughout the upper portion of the vault,
drift in the lower portions of the vault and
cranial base (Duterloo & Enlow, 1970;
Lieberman, 1996), and endochondral
growth within certain synchondroses (Figure
1). Antero-posterior and lateral NBC
growth occur through coronally-oriented
and sagittally-oriented sutures and synchondroses,
respectively. The role of basicranial
flexion is an additionally important, but
complex component of NBC expansion that
requires consideration. Ross & Ravosa
(1993) demonstrated that among anthropoid
non-human primates variations in
basicranial flexion are most probably
adaptations to accommodate increases in
brain size relative to cranial base length
(Biegert, 1963; Gould, 1977). Basicranial
flexion, however, remains constant (albeit
variable) and independent of endocranial
volume and cranial base length among the
Hominidae, suggesting that other processes
such as those listed above account for the
additional increases in cranial volume
1In cases of anencephaly, the neurocranium fails to
develop due to the absence of cerebral tissue and its
meningeal coverings (see Sperber, 1989).
      293
that characterize the genus Homo (Ross &
Henneberg, 1995).
The model presented here proposes that
variations in human NBC shape should,
therefore, arise primarily from elongation
and widening of the cranial base combined
with neurocranial growth in the lateral, posterior
and superior directions. These processes
are not independent because it is
likely that the cranial base and vault influence
each other’s growth, particularly in
early development. Studies of artificial vault
deformation clearly demonstrate some
effects of neurocranial growth on cranial
base shape: anterior–posterior head-binding
increases the breadth of the lateral portion
of the cranial base, circumferential headbinding
elongates the foramen magnum,
and both practices inhibit basicranial flexion
and alter the timing of synchondroseal
fusion (Anto´n, 1989; see also Moss, 1958;
McNeil & Newton, 1965; Plourde & Anto´n,

+ +
+ – +

(d) +
+ +
+ + + + + +

– – – – –

Figure 1. Model of integrated growth in the neurobasicranial (NBC) complex (derived from Enlow,
1990). (a) Brain expansion in midsagittal plane; (b) neurocranial and basicranial growth sites in composite
lateral and midsagittal view; (c) brain expansion in coronal plane; (d) neurocranial and basicranial growth
sites in posterior view. Open arrows indicate directions of neural expansion; closed arrows indicate sutural
and synchondroseal growth directions; + indicates sites of pericranial and endocranial bone deposition; 
indicates sites of pericranial and endocranial bone resorption. Expansion of the brain induces posteriorand
superior-, and to a lesser extent inferior- and anterior-directed tension in the neurocranium and
basicranium (a and c). The NBC expands in response to tension through intra-sutural and synchondroseal
growth (arrows in b and d), and through drift below the circumcranial reverse line (dark-shaded areas).
294 . .  ET AL.
1992). However, there are several reasons to
hypothesize that the basicranium exerts
more constraints on neurocranial growth
than vice versa. First, the initial morphologies
of endochondral bones, which
derive from segmented paraxial mesoderm,
may be less subject to epigenetic effects from
interactions with other tissues than those
of neural crest-derived intramembranous
bones (Hall, 1978; Jacobsen, 1993;
Thorogood, 1993). In addition, abnormalities
of cerebral shape and/or size such as
microcephaly and hydrocephaly tend to influence
the shape of the neurocranium
more than the basicranium (de Beer, 1937;
Weidenreich, 1941; Babineau & Kronman,
1969; David et al., 1990). An early onset
of hydrocephaly, for example, results in a
wide array of changes in neurocranial
shape, but mostly causes basicranial widening
(Richards & Anto´n, 1991). Moreover,
Howells (1969, 1973) has shown that variations
in basicranial breadth are the greatest
non-facial source of cranial variation among
modern human populations. Finally, it is
important to note that studies of the effects
of artificial cranial deformation demonstrate
that some alterations of the cranial vault
such as cradle-boarding among the Hopi
tend to have less pronounced effects on
the endochondrally-derived portions of the
basicranium (e.g., Kohn et al., 1995), while
other forms of artificial deformation such as
annular and anteroposterior deformations
produce similar magnitudes of shape
change in the vault, face, and basicranium
(Cheverud et al., 1992; Kohn et al., 1993;
Anto´n 1989, 1994). However, perturbations
of basicranial growth sometimes have profound
effects on neurocranial shape (Bu¨ tow,
1990). As noted above, anterior–posterior
head-binding can cause lateral expansion of
the cranial base (Anto´n, 1989; Cheverud
et al., 1992), primarily by widening the most
lateral portions of the cranial base around
the temporo-mandibular joint (Anto´n,
1989). In cases of artificial deformation,
however, the timing of the application of
external stresses to the growing cranium may
play a crucial role in whether the deformation
produces compensatory growth in the
face and basicranium as well as the vault.
Deformations applied during the first year of
life—when the neurocranium, basicranium,
and face are all growing at a rapid rate—
influence the growth of all of these regions;
deformations that begin later (after 2 to 4
years) have less potential to influence the
growth of the basicranium. Artificial deformation,
therefore, acts as a natural experiment
which alters the growth of some or
all of the components of the skull, depending
on its timing in ontogeny and the
specific portions of the skull that are developmentally
constrained due to the type of
With regard to the development of certain
aspects of cranial form, Green & Smith
(Green, 1990; Green & Smith, 1991; Smith
& Green, 1991) have proposed an alternative,
ontogenetically-based model for the
development of occipital buns and many of
the other cranial traits that distinguish
Neanderthals from modern humans. Green
& Smith (1991) hypothesize that the overall
cranial form of Neanderthals, including an
occipital bun, midfacial prognathism and a
strongly projecting supraorbital torus, all
result from accelerated growth of the
components of the basicranium. A test of
Green & Smith’s (1991) model would
require an examination of the growth rates
2Although Cheverud et al. (1992) and Kohn et al.
(1993) report that annular vault and fronto-occipital
vault deformation cause specific growth changes in
either the cranial base or vault, it is difficult to discern
whether these changes are directly responsible for ‘‘primary
evolutionary changes’’ in the vault, base, and/or
face (Cheverud et al. 1992: 343). This uncertainty
stems from the fact that application of cranial deformation
techniques involve apparatuses that apply strain
to both the neurocranium and basicranium simultaneously.
Therefore, at least in regard to annular vault
and fronto-occipital vault deformation techniques, it is
still unclear whether changes in the cranial base or vault
are responsible for the changes seen in the rest of the
      295
of the components of the basicranium in
juvenile Neanderthals and several ontogenetic
series of modern humans drawn from
populations that differ in their adult morphology.
The focus of the present study is
upon patterns of correlations among dimensions
in the adult cranium (which serve as a
record of total attained growth throughout
ontogeny), and thus does not address Green
& Smith’s (1991) hypothesis.
Another prediction of the model tested in
the present study is that the major dimensions
of the basicranium and neurocranium
exert an influence on facial growth. Like
the neurocranium, the face grows through
intramembranous ossification of neural
crest-derived tissue (facial prominences and
branchial arches) (Couly et al., 1993; Le
Douarin et al., 1993; Selleck et al., 1993).
It is widely appreciated that facial growth
is partially independent of the NBC, to a
large extent because much of it occurs in a
skeletal growth trajectory after the completion
of neural expansion (Moss & Young,
1960; Moore & Lavelle, 1974; Sirianni &
Swindler, 1979; Sirianni, 1985; Watts,
1985; Moyers, 1988). In humans, facial
growth is about 95% complete by 16–18
years, at least 10 years after the majority of
the neuro-basicranial complex has reached
adult size (Stamrud, 1959; Farkas et al.,
1992). Indeed, the genetic basis for lateroccurring
facial growth appears to be different
from that for the earlier expansion of the
neurocranium and basicranium (Cheverud,
1996). The basicranium and neurocranium,
however, may have some influence on the
growth of certain facial dimensions because
the upper face articulates with the anterior
cranial base and the anterior cranial fossa
and the mid-face articulates with the middle
cranial fossa. In particular, the upper and
middle portions of the face in humans grow
primarily from lateral drift and anterior displacement
around the ethmoid and in front
of the sphenoid (Sperber, 1989; Enlow,
1990). The upper face grows anteriorly and
inferiorly from the anterior cranial fossa
through drift, the middle face grows anteriorly
from the middle cranial fossa through
displacement; and the lower face drifts inferiorly
from the middle face and displaces
anteriorly from the back of the maxilla. As
Weidenreich (1941) suggested, the absolute
breadth of the neuro-basicranial complex
therefore probably constrains facial breadth.
This hypothesis receives some support from
studies of artificial cranial deformation.
Anto´n (1989, 1994), for example, has
shown that antero–posterior head-binding
during the first years of life causes not only a
wider neurocranium but also a concomitantly
wider face from additional growth in
the most lateral regions; conversely, circumferential
head-binding results in a narrower
neurocranium and face. In addition, cranial
base flexion may have some additional influences
on facial orientation such as the
degree of klinorhynchy (Enlow, 1968; Shea,
1985; Ravosa, 1991; Ross & Ravosa, 1993;
Anto´n, 1994; Ross & Henneberg, 1995),
but is not expected to influence other facial
Hypotheses to be tested
The above described developmental and
spatial interactions between the cranial base,
vault and face suggest that in adults, as a
final result of growth processes, the shape of
the cranial base may be correlated with the
shape of the neurocranium in several basic
ways. In particular, we propose several interrelated
hypotheses about predicted correlations
between selected basicranial, neurocranial,
and facial dimensions of adult
crania. Our first hypothesis regarding these
relationships concerns the relationship
between endocranial volume (ECV) and
basicranial form, under conditions in which
the major dimensions of the basicranium
(maximum breadth, maximum length, and
flexion) vary independently among adults
(an assumption which we test). The model
296 . .  ET AL.
proposes that the basicranium acts as a
platform upon which the brain expands, and
that when the basicranium ceases to grow,
its dimensions constrain the directions in
which the expanding brain and cranial vault
can grow. The model predicts that basicranial
width constrains the breadth of the
cranial vault, but that basicranial length (the
distance from basion to foramen caecum)
does not act as a strong constraint on the
growth of the brain posteriorly and any
associated elongation of the cranial vault. If
these relationships are true, we hypothesize
that breadth of the neurocranium correlates
positively with basicranial breadth and
endocranial volume, but is independent of
the length and degree of flexion of the
cranial base. In addition, the model predicts
fewer constraints on the length of the
neurocranium in adults, especially in the
posterior cranial fossa, than on its breadth.
Our second hypothesis, which follows
from the above model, is that variations in
the length of the neurocranium have a low
correlation with or are independent of
basicranial length and flexion, but correlate
positively with endocranial volume and
negatively with basicranial breadth. Third,
we hypothesize that variations in endocranial
volume relative to basicranial
breadth correlate positively with the height
of the neurocranium and with the degree of
posterior extension of the neurocranium in
the posterior cranial fossa. Another way of
stating this hypothesis is that given a large
brain and a narrow cranial base, the cranial
vault is likely to grow backward and upward
to accommodate the brain. As a result, the
degree of occipital bunning among recent
and fossil humans is predicted to be a
function of endocranial volume relative to
basicranial breadth, and is therefore
expected to occur more frequently in
large-brained, narrow-skulled individuals.
A second, related set of hypotheses concerns
the influence of the NBC on facial
growth as suggested by Weidenreich (1941),
Enlow & Bhatt (1984), Enlow (1990) and
others. Because the face grows downward
and forward from the cranial base, we
hypothesize that maximum upper facial
breadth in adult humans is constrained by
the breadth of the anterior cranial fossa and
that midfacial breadth is constrained by the
breadth of the middle cranial fossa. Enlow
(1990), Howells (1973), and a number of
researchers who have studied artificial vault
deformation (e.g., Cheverud et al., 1992;
Kohn et al., 1993; Anto´n, 1994) have also
suggested that NBC proportions influence
certain facial proportions. The basis for the
first hypothesis regarding facial dimensions
stems from the observation that growth of
the NBC, hence growth in midfacial
breadth, is complete long before the
majority of facial growth occurs. Therefore,
increases in facial size after the cessation of
the neural growth trajectory occur mostly as
anteriorly and inferiorly directed growth
(Moss & Young, 1960; Moore & Lavelle,
1974; Moyers, 1988; Farkas et al., 1992).
Enlow has specifically proposed that a
relationship exists between cranial form
and the prevalence of certain malocclusions
(illustrated in Figure 2). According to
Enlow (1990: 196–198), individuals with
absolutely narrow NBCs (primarily dolichocephalics)
tend to have more flexed
cranial bases, longer anterior cranial bases,
and narrower faces than individuals with
absolutely wider NBCs (primarily brachycephalics).
As a result, a second hypothesis
derives from Enlow’s (1990: 222–228) prediction
that individuals with absolutely
narrower NBCs have proportionately
narrower and antero-posteriorly longer
faces (leptoproscopy) than individuals with
wider NBCs who have proportionately
wider and antero-posteriorly shorter faces
Hypothesis testing
The above hypotheses can be tested using
comparisons of adult skulls or ontogenetic
      297
samples of skulls with different overall
cranial shapes. This paper employs the
former because of the difficulties of using
available ontogenetic data to distinguish
among the effects of the many diverse components
of the skulls that grow concurrently
in a similar trajectory. During the first 6
years of growth, the basicranium flexes and
the dimensions of the neurocranium and
basicranium all increase dramatically in size
in a common neural growth trajectory,
largely driven by the capsular functional
growth matrix of the expanding brain
(Moore & Lavelle, 1974; Ranly, 1988). Differences
in attained growth in adult crania
are relatively small compared to the disparity
in size between adults and neonates. As a
result, ontogenetic series of skulls inevitably
produce high correlations between endocranial
volume and all the components of
the NBC, making it difficult to factor out
autocorrelations that result from the tremendous
collinear growth that occurs in all
cranial dimensions in infancy and childhood
(for a discussion of these statistical problems,
see Sokal & Rohlf, 1995: 583–586).
Rather than focusing upon the correlations
between cranial dimensions during growth,
future studies may address these problems
by studying the timing of growth differences,
the correlations between growth rates, and
the relative timing of growth cessation
events between individuals (or populations)
that ultimately develop differently-shaped
crania. Relationships between elements of
the cranial base and neurocranial and facial
dimensions during ontogeny could be tested
using longitudinal data from many individuals
who differ in adult cranial shape,
or by comparing ontogenetic cross-sectional
samples from populations that differ in adult
cranial shape. In either case, the processes of
growth would not be expected to differ
fundamentally except perhaps in terms
of the slope and/or intercept values of
selected basicranial, neurocranial and facial
proportions throughout ontogeny. Investigation
of the relative timing of onset and
cessation of specific growth centers would
complete the picture of how differences in
adult morphology are achieved through
growth. Unfortunately, such longitudinal
Figure 2. Enlow’s (1990) model of differences in facial form between dolichocephalic and brachycephalic
individuals. According to this model, individuals with narrower NBCs will have proportionately narrower
and antero-posteriorly longer faces than individuals with broader NBCs. See text for details.
298 . .  ET AL.
and cross-sectional data are not available
currently, but hold much promise for future
The hypotheses presented above concerning
the influence of basicranial growth on
overall cranial shape can also be tested using
adult skulls from diverse populations that
sample a wide range of overall cranial
shapes. The adult skull is the final product
of ontogeny and represents the cumulative
product of the particular growth processes of
interest. In particular, this study uses a
pooled sample of males and females from
five craniometrically and geographically
diverse populations in order to examine as
wide a range of craniofacial forms as possible.
Although the inclusion of several populations
raises the possibility that some
proportion of the variation results from
inter-population differences (see below), we
emphasize that the hypotheses tested here
derive from a general model of craniofacial
growth that is not population or sex specific
and must apply to all craniofacial types. The
hypotheses we have proposed do not constitute
a test of the ontogenetic model itself,
but rather they serve as a test of the pattern
of correlations between the sizes of adult
structures expected from the model of
cranial ontogeny. For the ontogenetic model
to have any general validity, it must be able
to explain differences among as diverse an
array of cranial types as possible. As other
researchers (e.g., Howells, 1973; Lahr
1996) have shown, this kind of pooled
sample is necessary to examine a wider
range of craniofacial variation than is
present within single populations.
Materials and methods
The sample of recent Homo sapiens used in
this study comes from five geographically
and craniometrically diverse populations
from Australia, East Asia, Europe, North
Africa and sub-Saharan Africa (Table 1).
Roughly 20 adults whose M3s had fully
erupted were measured from each population.
An attempt was made to select equal
proportions of males and females from each
population by estimating sex using standard
sexually dimorphic characteristics (Bass,
1987: 81). As Figure 3 illustrates, these
skulls encompass a broad range of overall
cranial shapes: the Ashanti and Australian
individuals tend to be dolichocephalic, the
Chinese and Egyptian individuals tend to be
mesocephalic, and the Italian individuals
tend to be brachycephalic. Although the
pooled cranial sample is skewed towards
dolichocephalic individuals, this bias reflects
the prevalence of dolichocephaly among
human populations (Weidenreich, 1945;
Martin & Saller, 1956). A sample of
Pleistocene human skulls (archaic and anatomically
modern) which are substantially
complete and for which lateral radiographs
were available were also studied.
A series of measurements (listed in Table 2)
were taken on each cranium from external
landmarks and from radiographs. Summary
statistics for these measurements are provided
in Table 3. Exocranial linear dimensions
to the nearest 0·1 mm were taken
using Mitutuyo digital sliding or spreading
calipers. Lateral and superior–inferior radiographs
were taken of all specimens using an
ACOMA portable X-ray machine on
Kodak XTL-2 film. To minimize potential
distortion and parallax, care was taken to
orient the midsagittal plane of each cranium
parallel to the X-ray film and collimator for
the lateral radiographs and in the Frankfurt
horizontal for the supero-inferior radiographs.
Linear measurements of the radiographs
were taken to the nearest 0·1 mm
from tracings using Mitutuyo digital
sliding calipers. All linear radiograph
measurements were adjusted for size distortion
using a correction factor calculated as
the ratio of maximum cranial length
      299
Table 1 Samples used
Sample Taxon
Modern samples (S.D.s in parentheses)
n (m/f) ECV POB CI BI Location
Ashanti Recent Homo sapiens 18 (9/9) 1404·4 107·7 0·73 2·17 AMNH
(145·5) (4·5) (0·04) (0·4)
Australians Recent H. sapiens 21 (12/19) 1287·8 111·3 0·71 1·93 AMNH
(125·8) (5·3) (0·03) (0·67)
Chinese Recent H. sapiens 19 (10/9) 1496·7 116·7 0·79 1·66 AMNH
(92·3) (5·0) (0·04) (0·47)
Egyptians Recent H. sapiens 20 (10/10) 1342·4 108·7 0·75 2·10 PM
(119·1) (4·9) (0·03) (0·64)
Italians Recent H. sapiens 20 (11/9) 1350·3 118·0 0·82 1·53 PM
(151·8) (4·9) (0·03) (0·60)
Skull Taxon ECV* POB CI* BI Source of radiograph
Abri Pataud H. sapiens 1380 104 75·4 4 D. Lieberman (this study)
Skhul 4 H. sapiens 1554 117 71·8 3 B. Arensburg (personal communication)
Skhul 5 H. sapiens 1518 116 74·5 4 D. Lieberman (this study)
Cro Magnon 1 H. sapiens 1600 115 74·0 4 D. Lieberman (this study)
Obercassel 1 H. sapiens 1500 129 74·2 2 J. Weiner (NHM)
Obercassel 2 H. sapiens 1279 114 71·3 2 J. Weiner (NHM)
La Chapelle Homo neanderthalensis 1626 119 75·0 5 D. Lieberman (this study)
Monte Circeo H. neanderthalensis 1330 129 75·9 5 J. Weiner (NHM)
La Ferrassie 1 H. neanderthalensis 1681 128 76·1 5 D. Lieberman (this study)
La Quina 5 H. neanderthalensis 1350 119 — 5 J. Weiner (NHM)
*From Vandermeersch, 1981: 140–143.
Abbreviations in Table 1: ECV, endocranial volume; POB, bi-porionic breadth; CI, cranial index; BI, bunning index. Locations of the cranial collections:
AMNH, American Museum of Natural History; PM, Peabody Museum (Harvard University); NHM, Natural History Museum archives (courtesy of T.
300 . .  ET AL.
measured exocranially and maximum
cranial length measured from the radiograph.
Angular measurements of cranial
base flexion were made from tracings with a
protractor to the nearest degree. ECV was
measured in each cranium by filling the
neurocranial cavity with lentils through
the foramen magnum while shaking and
tapping the skull gently until no more
lentils could fit below the level of the
Cranial index
(n = 20)
(n = 20)
(n = 20)
(n = 20)
(n = 20)
Figure 3. Summary of variation in cranial index (XCB/GOL100) of recent human populations used in
this study.
Table 2 Linear and angular measurements used
Measurement ABBR. Definition
Neurocranial length GOL Chord distance from glabella to opisthocranion (Howells, 1973)
Max. cranial breadth XCB Maximum cranial breadth perpendicular to sagittal plane (Howells,
Max. basicranial breadth POB Bi-porionic breadth (Wood, 1991)
Upper facial breadth FMB Bi-frontomalare temporale breadth (Wood, 1991)
Mid-facial breadth JUB External malar breadth at jugale (Howells, 1973)
Lower facial breadth EPB External palate breadth at M1 (Wood, 1991)
Mid-facial breath MFB Bi-maxilofrontale breadth (Wood, 1991)
Neurocranial height BBH Chord distance from basion to bregma (Howells, 1973)
Facial height NPH Chord distance from nasion to prosthion (Howells, 1973)
Orbital height, left OBH Chord distance between upper and lower borders of orbit, perpendicular
to the long axis of orbit (Howells, 1973)
Cranial base lengthL BCL Chord distance from basion-sella plus sella-foramen caecum
Cranial base angleL CBA Angle between basion-sella and sella-foramen caecum
Ant. cranial base lengthL ACL Chord distance from sella-foramen caecum
Mid-facial lengthL MFL Minimum chord from PM plane* to nasion (Lieberman, in press)
Lower facial lengthL LFL Chord distance from anterior to posterior nasal spines
Ant. cranial fossa breadthS ACB Maximum anterior cranial fossa breadth anterior to clinoid processes
Mid. cranial fossa breadthS MCB Middle cranial fossa breadth at sella
LFrom lateral radiograph.
SFrom supero-inferior radiograph.
*The posterior maxillary (PM) plane, from the maxillary tuberosities to the anterior-most point where the
greater wings of the sphenoid intersect the planum sphenoideum (the junction of the anterior and middle cranial
fossa), is the boundary between the ethmomaxillary facial complex and the middle cranial fossa (Enlow, 1990).
      301
Table 3 Descriptive statistics of modern human sample (in mm, standard deviations in parentheses)
Ashanti 18 131·6 129·3 177·4 132·7 43·1 42·8 66·2 34·0 101·7 112·2 60·9 103·3 116·2 2·64
(4·7) (6·9) (7·1) (5·9) (2·8) (3·0) (4·9) (2·5) (4·4) (4·5) (3·8) (5·5) (5·9) (0·19)
Australians 21 132·4 125·4 176·0 129·8 45·0 43·6 63·8 33·3 106·0 114·2 61·2 103·1 116·0 2·63
(4·5) (6·0) (8·4) (4·9) (3·4) (2·1) (4·3) (1·8) (4·2) (4·7) (3·5) (4·5) (4·6) (0·16)
Chinese 19 132·8 137·3 175·5 132·8 42·8 39·7 70·5 34·5 103·7 114·3 60·5 109·3 125·5 2·90
(6·7) (4·2) (8·7) (6·6) (3·9) (2·7) (5·2) (2·2) (3·4) (3·9) (3·9) (4·5) (5·2) (0·21)
Egyptians 20 136·3 134·8 181·0 128·9 41·6 37·6 64·0 35·3 97·7 105·6 54·9 101·3 112·3 2·83
(3·0) (5·9) (6·3) (5·6) (4·5) (2·5) (4·6) (4·7) (3·8) (3·8) (3·1) (3·2) (5·0) (0·18)
Italians 20 135·7 142·3 174·2 130·3 43·6 39·0 62·3 34·3 102·1 112·8 55·5 107·4 119·6 2·93
(2·8) (6·2) (8·2) (6·5) (4·2) (3·7) (4·6) (5·1) (5·6) (7·8) (5·5) (7·2) (7·4) (0·24)
Combined 98 133·8 133·7 176·8 130·8 43·3 40·5 65·2 34·3 102·3 111·7 58·5 104·8 117·8 2·79
(4·8) (8·4) (7·7) (5·9) (3·9) (3·6) (5·4) (3·5) (5·1) (6·1) (4·9) (5·8) (7·0) (0·23)
See Table 2 for measurement details and abbreviations; FI, facial index (JUB/MFL100).
302 . .  ET AL.
foramen magnum. The lentils were then
emptied via a funnel into a graduated cylinder
which was shaken until they had subsided
completely, and the volume read to
the nearest milliliter. In order to make the
volumetric measurements similar in scale to
the linear dimensions of the cranium, the
measurement of ECV used in the analysis is
calculated as the cube root of the measured
cranial capacity.
Although widely considered to be a continuously
variable feature, occipital bunning
is difficult to quantify because it combines
projection of both the internal and outer
tables of the occipital inferior to lambda and
superior to the internal occipital protruberance
(IOP), the attachment site of the tentorium
cerebelli (Ducros, 1967; Trinkaus
& LeMay, 1982; Lieberman, 1995). Appraisals
of occipital bunning probably often
rely upon subjective, visual examination of
the amount of external projection of the
occiput. Our definition, based on Trinkaus
& LeMay (1982), primarily emphasizes the
internal curvature of the occipital bone,
which more closely reflects the shape of
the brain. Following Lahr’s (1994, 1996)
approach of dividing continuously varying
aspects of cranial morphology into discrete
categories, the degree of occipital bunning
was quantified in each individual from lateral
radiographs using a system of grades,
reflecting examples of minimum, intermediate
and maximum degrees of development
selected from the total sample studied
(including Neanderthals). Five grades (illustrated
in Figure 4) were defined using the
contours of both internal and outer tables
of the occipital from lateral radiographs as
follows: 1, no bunning, with very little
posterior projection of the occipital tables
above the IOP and below lambda; 2, minor
bunning, with slight posterior projection of
the internal table of the occipital above
the IOP and below lambda; 3, moderate
bunning, with a significantly more concave
internal occipital table above the IOP and
below lambda; 4, marked bunning, with a
Lambda Lambda
Figure 4. Grades used to evaluate degree of occipital bunning in individuals from lateral radiographs. See
text for details. Occipital projection is evaluated primarily from the internal occipital table relative to the
internal occipital protruberance (IOP) and to lambda (L). Note that grades 3 and 4 differ solely in terms
of the morphology of the outer table. Most previous appraisals of bun development have generally been
made upon the external morphology of the occipital rather than the internal contour, as used here.
      303
similar degree of internal occipital concavity
to grade 3, but in which the external table is
also substantially more developed above the
IOP and below lambda; 5, extreme bunning,
with a highly concave internal occipital table
above the IOP and below lambda, combined
with a thick and highly convex outer table.
Note that grades 3 and 4 differ solely in
terms of the morphology of the outer table.
An important detail with respect to occipital
bunning concerns the fossil specimens
that could be included in the analysis. Most
of the early modern humans that have pronounced
occipital buns come from central
European Aurignacian or Gravettian sites
such as Mladec, Zlaty Kun, Stetten,
Prˇedmostı´, Brno, Dolnı´ Veˇstonice, and
Pavlov (Jelı´nek, 1969; Smith, 1982, 1984;
Vlcˇek, 1991). The crania from Prˇedmostı´
were destroyed in World War II, and other
specimens lack a cranial base, limiting their
value for this study. Similarly, we were
unable to obtain radiographs of some of the
more complete Central European specimens,
and the western European early modern
specimen that has the most pronounced
occipital protuberance, Cro-Magnon 3, was
not included because its cranial base is not
preserved. Thus our sample of early modern
humans does not include many of the specimens
in which external bunning is most
pronounced and it does not include the
Mladec specimens, which play a crucial role
in discussions of possible continuity from
Neanderthals to modern humans in Central
Europe (e.g., Jelı´nek, 1969; Smith, 1984;
Frayer, 1986; Wolpoff, 1996). Thus the
sample of measurements and radiographs of
early modern humans we obtained can only
serve as a preliminary test of the hypothesis
that early modern humans and Neanderthals
developed occipital buns in different
Statistical analysis
All measurements were entered and analyzed
using Statview 4.5 (Abacus Concepts,
Berkeley, CA, U.S.A.) and Systat 5.2
(Systat Inc., Evanston, IL, U.S.A.). The
accuracy of the linear, angular and volumetric
measurements were tested by taking each
measurement five times on the same skull.
Average measurement error was 1.4%. The
cube-root of ECV was used in order to
compare linear and volumetric measurements.
Normality was tested for each variable
using the Lilliefors test (Lilliefors,
1967). In order to examine the effects of
overall cranial size, a geometric mean
(GGM) of ten diverse craniofacial dimensions
was computed as the tenth root of the
product of the following measurements:
ECV, maximum cranial breadth (XCB),
upper facial breadth (FMB), mid-facial
breadth (JUB), neurocranial length (GOL),
facial height (NPH), orbital height (OBH),
lower facial length (LFL), maximum
basicranial (bi-porionic) breadth (POB),
and basion-bregma height (BBH); a geometric
mean of overall facial size (FGM)
was calculated as the fifth root of the product
of facial height (NPH), orbital height
(OBH), midfacial breadth (MFB), lower
facial breadth (EPB), and lower facial length
Because the majority of the variables
examined in this study are normally distributed
(see below), predicted relationships
among craniofacial dimensions were examined
in the pooled modern human sample
primarily using Pearson correlation coefficients,
and using partial correlation analysis
in order to hold certain variables (which
serve as a proxy for overall cranial size)
constant. Linear regression and stepwise
multiple regression analyses are also used
to estimate the proportion of variance
explained by specific variables. The
strengths of the correlations among categorical
data (e.g., occipital bunning) and any
non-normally distributed variables are
examined with Spearman rank correlation
analysis. Significance testing of Pearson
correlation coefficients was determined
304 . .  ET AL.
using Fisher r–z test; significance of partial
correlations was determined using the significance
of partial regression coefficients.
Basicranial–neurocranial interactions
The first hypothesis regarding relationships
between the basicranium and neurocranium
predicts that the length and breadth of
the cranial base should be independent, and
that the interaction of endocranial volume
with these two basicranial dimensions
potentially influences the major dimensions
of the neurocranium. Correlation and partial
correlations in Table 4 indicate that
there is a moderate, significant correlation
between cranial base length (BCL) and
POB, but that these two dimensions are
entirely independent (r=0·026) when one
holds CBA, ECV and GGM constant using
partial correlation analysis. These results
also indicate that variations in ECV do not
correlate with variation in BCL and POB
when one holds overall cranial size constant,
and confirms the hypothesis that variation in
CBA among adult humans is independent of
ECV, overall cranial size, and other cranial
base dimensions (Ross & Henneberg,
1995). These data, therefore, clearly support
the hypothesis that attained growth in
the breadth, length and flexion of the cranial
base result from independent processes in
the sense that they do not appear to affect
each other’s size.
A second hypothesis is that both POB and
ECV influence the maximum breadth of the
neurocranium, but that maximum neurocranial
breadth is expected to be independent
of cranial base length and flexion. The pairwise
correlations show that there is a weak
but highly significant (r=0·283, P<0·001)
association between basicranial length and
maximum neurocranial breadth, while the
correlation between maximum cranial
breadth and cranial base flexion is both low
and non-significant (Table 5). If considered
in isolation, the significant correlation
between XCB and BCL might falsify the
first hypothesis. However, the hypothesis is
supported more strongly by partial correlation
analysis (Table 5), which shows that
XCB correlates moderately and significantly
with POB (r=0·440, P<0·05) and ECV
(r=0·358, P<0·05), and that these correlations
remain moderate when one holds
BCL, CBA and GGM constant. In addition,
BCL, CBA and the GGM measurements all
have extremely low partial correlations with
XCB. In other words, basicranial breadth
and endocranial volume appear to be major
influences on neurocranial breadth, holding
overall cranial size constant. According to a
stepwise multiple regression, POB and ECV
together account for approximately 56% of
the varation in XCB (P<0·001) in the recent
human sample studied here.
Given the independence between cranial
base breadth and length and the observation
that POB and ECV remain associated with
Table 4 Correlation (top) and partial correlation (bottom) analysis of major basicranial and neurocranial
POB — 0·437‡ 0·005 0·409‡ 0·644
BCL 0·026 — 0·031 0·453‡ 0·697‡
CBA 0·041 0·096 — 0·025 0·053
ECV 0·041 0·027 0·018 — 0·670‡
GGM 0·476* 0·538* 0·104 0·505‡ —
*P<0·05; †P<0·01; ‡P<0·001.
      305
variation in XCB when other factors (e.g.,
GGM as a measure of overall cranial size)
are held constant, the developmental model
tested here predicts that there should be
fewer constraints on neurocranial length and
height, especially in the posterior cranial
fossa. In particular, the second hypothesis
predicts that GOL should be independent
with or have a low correlation with BCL and
CBA, but should be highly correlated with
ECV, holding overall cranial size and the
other variables constant. This hypothesis is
tested using correlation and partial correlation
analysis in Table 6. As these analyses
indicate, GOL is moderately and significantly
(P<0·001) correlated with BCL,
ECV and GGM, which would seem to disprove
the hypothesis. However, when the
interactions among these variables are
removed using partial correlation, GOL is
independent of CBA, BCL and GGM, and
has a moderate but significant (partial
r=0·414; P<0·05) partial correlations with
ECV. In other words, the second hypothesis
is supported because neither cranial base
length nor flexion has much influence on
total neurocranial length independent of
other factors.
Since cranial base breadth apparently
constrains neurocranial breadth, cranial
base length is independent of neurocranial
length, and ECV significantly affects both
neurocranial dimensions, then it follows that
the developmental interactions between
brain volume and cranial base breadth
should influence the length and height of the
neurocranium more than its breadth. This
hypothesis is tested in Table 7 using correlation
and partial correlation analysis. While
the Pearson correlation coefficients between
the ratio of ECV/POB and GOL, XCB,
CBA and BBH are all quite low, there is a
moderate but significant partial correlation
between the ratio of ECV/POB and GOL
(partial r=0·400), holding the other dimensions
including GGM constant. In other
Table 5 Correlation (top) and partial correlation (bottom) analysis basicranial and neurocranial
breadth with endocranial volume, basicranial flexion, and overall cranial size
XCB — 0·592‡ 0·541‡ 0·283‡ 0·139 0·516‡
POB 0·440* — 0·409 0·427‡ 0·005 0·644‡
ECV 0·358* 0·192 — 0·453‡ 0·025 0·670‡
BCL 0·111 0·025 0·014 — 0·031 0·697‡
CBA 0·165 0·109 0·076 0·076 — 0·053
GGM 0·030 0·414* 0·461* 0·538* 0·097 —
*P<0·05; †P<0·01; ‡P<0·001.
Table 6 Correlation (top) and partial correlation (bottom) analysis of neurocranial length with
basicranial length, basicranial flexion, endocranial volume, and overall cranial size
GOL — 0·397‡ 0·137 0·605‡ 0·520‡
BCL 0·090 — 0·031 0·453* 0·697‡
CBA 0·154 0·107 — 0·025 0·053
ECV 0·414* 0·061 0·079 — 0·670‡
GGM 0·100 0·583‡ 0·079 0·459‡ —
*P<0·05; †P<0·01; ‡P<0·001.
306 . .  ET AL.
words, individuals with larger brains relative
to basicranial breadth tend to have slightly
longer cranial vaults, controlling for other
factors. It is important to note, however,
that in a multiple stepwise regression, variations
in ECV and POB explain only 35% of
the variation in GOL (P<0·001) and 38% of
the variation in BBH (P<0·001), indicating
that other factors probably related to brain
shape and overall cranial size (e.g., the dural
bands) have more dominant influences on
these dimensions. The Pearson and partial
correlations reveal that individuals with a
high ratio of ECV/POB also tend to have a
cranium with a relatively small GGM.
These results support the hypothesis that
the cranial vault tends to expand vertically
and especially posteriorly to a greater extent
in individuals with a large brain relative to
cranial base breadth. The interactions
between brain volume and cranial base
breadth that apparently influence neurocranial
length may, therefore, explain some
of the variation in occipital projection or
‘‘bunning’’ in recent and Pleistocene
humans. Among the recent human sample
studied here, posterior projection of the
occipital is generally modest: 87% were
scored as having no bunning (type I) or only
slight bunning (type II), and only 13% were
scored as having moderate bunning (type
III). However, the Spearman rank correlation
between the degree of occipital
projection and the ratio of ECV/POB is
moderate and highly significant (=0·600;
P<0·001), and thus supports the hypothesis
that posterior projection of the occipital is
partially a function of having a large brain
relative to cranial base area. As predicted by
Lieberman (1995), the degree of bunning
also has a moderate negative Spearman
rank correlation (=0·428, P<0·001)
with the cranial index, indicating that
dolichocephalic individuals are more likely
to have posteriorly-projecting occipitals
than brachycephalic individuals. As noted
by Trinkaus & LeMay (1982), posterior
projection of the occipital in these humans
was always above the tentorium cerebelli,
suggesting that the expansion was probably
a result of local growth effects exerted by the
occipital lobe on the occipital squama.
When the Pleistocene sample of recent
modern humans, who tend to have large
brains and narrow cranial bases, is added to
the analysis, the relationship between occipital
bunning and between brain size is
strengthened (Figure 5). However, it is clear
from Figure 5 that Neanderthals do not fit
the early modern human pattern, primarily
because they have somewhat wider cranial
bases relative to endocranial volume [which,
with the small sample size used, does not
reach significance (P=0·16)], and marked
occipital buns (Grade 5). Other differences
in bunning are also evident from Figure 6,
which compares lateral radiographs of
the occipital in several Neanderthals and
Pleistocene modern humans. As Figure 6
shows, while some early modern humans
Table 7 Correlation (top) and partial correlation (bottom) analysis of the ratio of endocranial volume
to basicranial breadth with major neurocranial dimensions
ECV/POB — 0·225* 0·254 0·075 0·004 0·233
GOL 0·400* — 0·247* 0·432‡ 0·137 0·520‡
XCB 0·124 0·025 — 0·240* 0·139 0·516‡
BBH 0·268* 0·005 0·106 — 0·012 0·657‡
CBA 0·012 0·129 0·127 0·024 — 0·053
GGM 0·408‡ 0·432‡ 0·375‡ 0·603‡ 0·053 —
*P<0·05; †P<0·01; ‡P<0·001.
      307
have posteriorly projecting occipitals relative
to the IOP and to lambda, none have the
degree of internal table concavity that is
typical of Neanderthals. Viewed externally,
however, many of these early modern
human fossils have marked ‘‘buns’’ because
of the thickness of the cranial vault in this
region (Lieberman, 1996). It is worth reiterating
that, for developmental reasons,
the definition of occipital buns used here
places the greatest emphasis upon the curvature
of the internal table of the occipital.
The results support Trinkaus & LeMay’s
(1982) hypothesis that other developmental
factors—possibly related to differences in
the timing of posterior cerebral growth relative
to the growth of the cranial vault
bones—apparently account for the extreme
degree of posterior projection of the occipital
in the Neanderthals. Therefore, it
appears that the externally visible similarities
in occipital form between large-brained
dolichocephalic humans and Neanderthals
may not be entirely homologous in a developmental
sense. Indeed, Smith (1984) has
noted that the form of occipital buns differs
in Neanderthals and robust early modern
humans, leading him to describe the occipital
protuberances of the early moderns as
‘‘hemi-buns.’’ If these morphologically
divergent structures are not developmentally
homologous, then the presence of occipital
buns in Neanderthals and post-Neanderthal
Europeans does not necessarily indicate
genetic continuity.
The above results regarding the relationship
between bunning and brain size relative
to cranial base width must be interpreted
with some caution, however, given problems
with the sample studied here. Our sample of
early modern humans does not include
Eastern European specimens such as the
Mladec, Prˇedmostı´, Dolnı´ Veˇstonice, and
Pavlov which have some of the largest
occipital protuberances of early modern
crania. In addition, the early modern sample
does not include Cro-Magnon 3, the western
European early modern cranium with
the greatest degree of bun development,
because the specimen lacks its cranial
base. Furthermore, although bi-porionic
breadths (POB) have not been published for
the Mladec crania, Frayer (1986) reports
measurements for bi-auricular breadth
(which is usually a few millimeters larger
than POB) for Mladec 1, 2, and 5 (128·8,
130·4, and 150·0 mm, respectively). These
bi-auricular dimensions are remarkably
0.08 0.115
Bun index
0.085 0.09 0.095 0.100 0.11
Early modern H. sapiens
Figure 5. Variation in occipital bunning plotted against endocranial volume (cube root) divided by
bi-porionic breadth. H. sapiens, but not Neanderthals with larger endocranial volumes relative to POB,
tend to have significantly more posterior projection of the occipital.
308 . .  ET AL.
large, comparable with, or even larger than,
the corresponding POB measurements in
the Neanderthals sampled here. Therefore,
the argument that early modern humans
developed buns upon narrow cranial bases,
while Neanderthals developed their buns
upon wide cranial bases probably cannot be
extended to the early modern crania from
Neuro-basicranial–facial interactions
The second set of hypotheses tested here
attempts to relate the potential developmental
influence of neuro-basicranial complex
(NBC) breadth on certain facial dimensions
as predicted by Weidenreich (1941) and
Enlow (1990). According to the above
model, the breadth of the upper face is
predicted to be constrained by the breadth
of the upper cranial fossa, and the breadth of
the midface is predicted to be constrained by
the breadth of the middle cranial fossa.
Correlation analyses provide some support
for this hypothesis. Among the recent
human sample, Pearson correlation coefficients
between upper facial breadth and
anterior cranial fossa breadth (r=0·532,
P<0·001) and between midfacial breadth
and middle cranial fossa breadth (r=0·490,
P<0·001) are moderate in strength and
highly significant. However, Enlow’s (1990)
hypothesis that NBC dimensions influence
facial proportions receives only partial support.
Among the sample studied here, there
is no significant correlation between XCB
or the cranial index (CI) with either CBA
or anterior cranial base length (ACL), as
the model predicts. Nevertheless, there
is a moderate and significant correlation
(r=0·491, P<0·001) between XCB and the
facial index (FI)—the ratio of facial breadth
(JUB) to facial length (MFL)—which yields
a partial correlation coefficient of r=0·397
when overall facial size (FGM) is held constant.
Moreover, there is also a significant,
moderate correlation between the cranial
and facial indices (r=0·492, P<0·001),
which yields a partial correlation of r=0·499
when FGM is held constant. In other words,
Enlow’s observation that individuals with
absolutely and relatively narrower neurobasicranial
complexes (NBCs) tend to have
proportionately longer (antero-posteriorly)
and narrower faces than individuals with
wider NBCs is partially supported, but the
trend is weak and is largely the result of
variation between populations rather than
within them (see Figure 7). In particular,
Italian and Chinese individuals tend to have
relatively high values for the facial index and
a broad neurocranium, while Australian
and, to a lesser extent, Ashanti individuals
Figure 6. Lateral radiographs of posterior cranium in
Abri Pataud (a), Cro-Magnon I (b), La Chapelle aux
Saints (c) and La Ferrassie I (d). Internal occipital
protruberance, X; Lambda, L.
      309
tend to have a low facial index and a narrow
neurocranium. Variations in overall NBC
shape (as expressed by the cranial index)
account for approximately only 25% of the
variation in the facial index, highlighting the
high degree of variability of facial form in
relation to cranial form between and, to a
lesser extent, within, recent human populations
that must be explained by other factors
unrelated to the NBC.
The above results support the hypothesis
that certain dimensions of the cranial base,
mostly bi-porionic breadth, do have some
influence on the shape of the NBC, as noted
by Howells (1969, 1973). In the adult cranium,
variations in the breadth, length and
flexion of the cranial base are independent
of each other, and POB affects overall NBC
shape to some extent by influencing the
breadth of the neurocranium. However, the
effects of POB on neurocranial shape mostly
occur as the result of interactions with brain
size, and the correlations between POB and
most NBC dimensions (e.g., cranial vault
length and height) are only moderate, indicating
that other factors such as overall
cranial size have substantial effects. In
addition, it is important to stress that variations
in basicranial length and flexion appear
to have no significant influence on most
aspects of craniofacial shape independent of
other factors. These results do not necessarily
mean that CBA and CBL do not affect
NBC shape, but their effects are apparently
more regional. Lieberman (1998), for
example, has shown that in ontogenetic
samples of humans and chimpanzees
the length of the sphenoid body affects the
degree of facial projection relative to the
anterior cranial fossa, which in turn affects
on other aspects of cranial shape.
In addition, Weidenreich’s (1941) and
Enlow’s (1990) hypotheses that facial shape
is influenced by cranial base and neurocranial
dimensions are only weakly supported.
As the above results indicate,
narrow-skulled individuals tend to have narrower
faces than wider skulled individuals.
In addition, narrow-faced (leptoproscopic)
individuals tend to have antero-posteriorly
longer faces relative to facial breadth than
wide-faced (euryproscopic) individuals, but
these correlations account for only about
25% of the variation in the facial index
among adults. In other words, Enlow’s
(1990) prediction that individuals with
absolutely and relatively narrower NBCs
110 155
Maximum neurocranial breadth (XCB), mm
Facial index
115 120 125 130 140
145 150
Figure 7. Bivariate scattergram of facial index (JUB/MFL) versus maximum NBC breadth (XCB) in
recent modern human sample.
310 . .  ET AL.
should have proportionately longer (anteroposteriorly)
and narrower faces than individuals
with wider NBCs is only a tendency and
not a strong relationship. These results,
which suggest that most aspects of variation
in facial shape are independent of NBC
dimensions, should not be surprising given
the very different ontogenetic trajectories
of facial and NBC growth, and their
contrasting modes of growth.
These results need to be tested further
using ontogenetic samples from different
populations, and using large samples from
single populations. However, they confirm
the tremendous amount of integration that
occurs between the cranial base, the neurocranium
and the face during growth (see
also, Cheverud, 1982, 1996; Kohn et al.,
1993). Although it is tempting to consider
the neurocranium and basicranium to be
separate regions by virtue of their distinct
embryological origins, their dimensions
exhibit considerable intercorrelation in the
adult skull largely because of the unifying
capsular functional growth matrix of the
brain. One consequence of these integrative
processes is that there are few aspects of
neurocranial shape which are independent
of brain size and basicranial dimensions
such as POB. The phylogenetic implications
of these results are sobering because independence
is one of the major criteria for
characters in both phenetic and cladistic
analyses (Sokal & Sneath, 1963; Shaffer
et al., 1991; Lieberman, 1999). Characters
add phylogenetic information to an analysis
only to the extent that they are independent
of other characters; moreover, intercorrelated
characters will tend to bias phylogenetic
inferences incorrectly if they are
homoplastic or otherwise poor indicators
of ancestry and descent. Consequently,
measurements of overall cranial shape and
size are unlikely to be good characters for
either cladistic or phenetic analyses because
they result from processes of integration that
are likely to obscure independent, heritable
units of information (Bookstein, 1994;
Lieberman, 1999). For example, the above
results indicate that variations in overall
neurocranial form (dolichocephaly vs.
brachycephaly) result in part from variations
in basicranial breadth, but also derive
from interactions with brain size and other
factors that influence overall cranial size
such as long-term adaptations to climate
(Guglielmino-Matessi et al., 1979; Beals
et al., 1984).
Following Cheverud (1996), it may be
more sensible to identify characters for
phylogenetic analysis that comprise several
highly intercorrelated dimensions that may
evolve together as integrated morphometric
units. It is also possible that other basicranial
characters may prove to be more
useful for phylogenetic analysis, but most of
these characters are likely to be local,
specific features whose growth and morphology
are largely independent of the
effects of brain growth and other dominant
functional matrices in the skull. As
Cheverud (1982, 1996) has shown, the
independence of such characters needs to be
tested using techniques such as cluster
analysis in order to examine the intensity of
statistical associations among characters
from functionally and developmentally
interdependent sets. Defining and recognizing
such characters is a challenge, however,
given the prevalence of epigenetic influences
on cranial growth. Consider, for example,
just a few of the epigenetic stimuli that affect
facial growth and integration. Intramembranous
bone growth around the oropharynx
and nasopharynx, which comprise much
of the middle and lower portions of the
face, are induced to a large extent by
air-flow resistance (Linder-Aronson, 1979;
Principato & Wolff, 1985; Franciscus &
Trinkaus, 1988; Cooper, 1989; Warren
et al., 1992; Franciscus, 1995). In addition,
mechanical strains from generating and
resisting masticatory force influence lower
and midfacial growth along a complex strain
      311
gradient, with more dominant effects occurring
in the mandible and zygomatic arches,
and lesser effects occurring further away
from the teeth or muscle attachment
sites (e.g., Carlson & Van Gerven, 1977;
Corruccini & Beecher, 1982; Kiliaridis
et al., 1986; Hylander, 1988; Hylander &
Johnson, 1992; Herring, 1993; Bouvier &
Hylander, 1997). Moreover, much of the
early growth of the orbital cavity (which
comprises parts of seven bones) is stimulated
by expansion of the eyeballs, but the
subsequent infero-lateral expansion of the
orbital cavities is a consequence of other
processes of facial growth (Enlow, 1990;
Denis et al., 1993).
Although this study does not directly test
the hypothesis that cranial base dimensions,
by virtue of their growth, are likely to be
better sources of phylogenetic information
than facial or neurocranial dimensions, these
results do highlight the utility of considering
how the pattern of correlations between the
cranial base and the brain contribute to
other important morphological variations.
In particular, these data clearly indicate that
the degree of posterior projection of the
occipital bone (bunning) may to some
extent be influenced by having a large brain
on a relatively narrow cranial base. Posterior
projection of the occipital lobe relative to the
internal occipital protruberance and to
lambda is more marked in large-brained,
dolichocephalic individuals, as predicted by
Trinkaus & LeMay (1982) and Lieberman
(1995). However, these interactions do not
explain the thickening of outer table that
results in Grade 4 occipital buns among
some robust early modern humans and the
extreme Grade 5 buns in Neanderthals.
These aspects of cranial vault thickness,
common to all Pleistocene populations, have
a different etiology, most probably a function
of systemic cranial and/or skeletal bone
growth (Lieberman, 1996). In addition, the
above results tentatively indicate that occipital
bunning in Neanderthals, who have wide
cranial bases relative to endocranial volume,
must be accounted for by other factors perhaps
related to the timing of brain growth
relative to basicranial growth, and thus may
not be entirely homologous with the morphology
occasionally evident in anatomically
modern H. sapiens. Caution on this last
point is required, however, because the
sample of early modern human analyzed
herein does not include many of the specimens
with the largest occipital buns (at least
as viewed externally). Further testing of this
hypothesis needs to include the early modern
crania from the early Aurignacian site of
Mladec, which resemble Neanderthals in
having distinct occipital protuberances and
wide cranial bases.
We thank S. Anto´n, J. Cheverud, W. W.
Howells, R. McCarthy, R. J. Smith, B.
Wood, D. Strait, F. Smith, and two anonymous
reviewers for comments on various
drafts, and we also thank R. McCarthy and
R. Bernstein for help in acquiring the data.
For assistance and access to collections,
we thank D. Pilbeam, Peabody Museum
(Harvard); A. W. Crompton, Museum of
Comparative Zoology; and I. Tattersall and
G. Sawyer, Department of Anthropology,
American Museum of Natural History. This
research was supported in part by funding
from grants from The Boise Fund (to
KMM) and the L.S.B. Leakey Foundation.
Anto´n, S. C. (1989). Intentional cranial vault deformation
and induced changes of the cranial base
and face. Am. J. phys. Anthrop. 79, 253–267.
Anto´n, S. C. (1994). Mechanical and other perspectives
on Neandertal craniofacial morphology. In
(R. S. Corruccini & R. L. Ciochon, Eds) Integrative
Paths to the Past: Palaeoanthropological Advances in
Honor of F. Clark Howell, pp. 677–695. Englewood
Cliffs, NJ: Prentice Hall.
Babineau, T. A. & Kronman, J. H. (1969). A cephalometric
evaluation of the cranial base in microcephaly.
Angle Orthod. 39, 57–63.
312 . .  ET AL.
Bass, W. M. (1987). Human Osteology: A Laboratory
and Field Maual, 3rd edn. Columbia, MO: Missouri
Archaeological Society.
Beals, K. L., Smith, C. L. & Dodd, S. M. (1984). Brain
size, cranial morphology, climate, and time
machines. Curr. Anthrop. 25, 301–330.
Biegert, J. (1963). The evaluation of characters of the
skull, hands and feet for primate taxonomy. In (S. L.
Washburn, Ed.) Classification and Human Evolution,
pp. 116–145. Chicago: Aldine.
Bookstein, F. L. (1994). Can biometrical shape be a
homologous character? In (B. K. Hall, Ed.) Homology:
The Hierarchical Basis of Comparative Biology,
pp. 197–227. San Diego: Academic Press.
Bosma, J. F. (1976). Development of the Basicranium.
Bethesda. Maryland: National Institutes of Health.
Bouvier, M. & Hylander, W. L. (1997). The function
of secondary osteonal bone: mechanical or metabolic?
Arch. Oral Biol. 41, 941–950.
Bu¨ tow, K-W. (1990). Craniofacial growth disturbance
after skull base and associated suture synostoses in
the newborn Chacma baboon: a preliminary report.
Cleft Palate J. 27, 241–252.
Carlson, D. S. & Van Gerven, D. P. (1977). Masticatory
function and Post-Pleistocene evolution in
Nubia. Am. J. phys. Anthrop. 46, 495–506.
Cheverud, J. M. (1982). Phenotypic, genetic, and
environmental morphological integration in the
cranium. Evol. 36, 499–516.
Cheverud, J. M. (1996). Developmental integration
and the evolution of pleiotropy. Am. Zool. 36, 44–50.
Cheverud, J. M., Kohn, L. A., Konigsberg, L. W. &
Leigh, S. R. (1992). Effects of fronto-occipital artificial
cranial vault modification on the cranial base and
face. Am. J. phys. Anthrop. 88, 323–345.
Cooper, B. C. (1989). Nasorespiratory function and
orofacial development. Otolaryng. Clin. N. Am. 22,
Corruccini, R. S. & Beecher, R. M. (1982). Occlusal
variation related to soft diet in a nonhuman primate.
Science 218, 74–76.
Couly, G. F., Coltey, P. M. & Le Douarin, N. M.
(1993). The triple origin of skull in higher
vertebrates. Develop. 117, 409–429.
David, D. J., Hemmy, D. C. & Cooter, R. D. (1990).
Craniofacial Deformities: Atlas of Three Dimensional
Reconstruction from Computed Tomography. New York:
de Beer, G. R. (1937). The Development of the Vertebrate
Skull. Oxford: Oxford University Press.
Denis, D., Faure, F., Volot, F., Scheiner, C., Boubli,
L., De´zard, X. & Saracco, J.-B. (1993). Ocular
growth in the fetus. 2. Comparative study of the
growth of the globe and the orbit and the
parameters of fetal growth. Opthalmologica 207, 125–
Ducros, A. (1967). Le chignon occipital, mesure sur le
squellete. L’Anthropologie 71, 75–96.
Duterloo, H. S. & Enlow, D. H. (1970). A comparative
study of cranial growth in Homo and Macaca. Am. J.
Anat. 127, 357–368.
Enlow, D. H. (1968). The Human Face: An Account of
the Postnatal Growth and Development of the Craniofacial
Skeleton. New York: Harper and Row.
Enlow, D. H. (1990). Facial Growth, 3rd edn.
Philadelphia: Saunders.
Enlow, D. H. & Bhatt, M. K. (1984). Facial
morphology variations associated with headform
variations. J. Charles Tweed Found. 12, 21–23.
Farkas, L. G., Posnick, J. C. & Hreczko, T. M. (1992).
Anthropometric growth study of the head. Cleft
Palate Craniofac. J. 29, 303–318.
Franciscus, R. G. (1995). Later Pleistocene nasofacial
variation in Western Eurasia and Africa and modern
human origins. Ph.D. Dissertation, University of
New Mexico.
Franciscus, R. G. & Trinkaus, E. (1988). Nasal morphology
and the emergence of Homo erectus. Am. J.
phys. Anthrop. 7, 517–527.
Frayer, D. W. (1986). Cranial variation at Mladec and
the relationship between Mousterian and Upper
Paleolithic hominids. Anthropos (Brno) 23, 243–256.
Frayer, D. W. (1992a). The persistence of Neanderthal
features in post-Neanderthal Europeans. In (G.
Bra¨uer & F. H. Smith, Eds) Continuity or Replacement:
Controversies in Homo sapiens Evolution, pp.
177–188. Rotterdam: A. A. Balkema.
Frayer, D. W. (1992b). Evolution at the European
edge: Neanderthal and Upper Paleolithic relationships.
Pre´histoire Europe´enne 2, 9–69.
Frayer, D. W., Wolpoff, M. H., Thorne, A. G., Smith,
F. H. & Pope, G. (1993). Theories of modern human
origins: the palaeontological test. Am. Anthrop. 95,
Friede, H. (1981). Normal development and growth of
the human neurocranium and cranial base. Scand. J.
Plas. Recon. Sur. 15, 163–169.
Gould, S. J. (1977). Ontogeny and Phylogeny.
Cambridge: Harvard University Press.
Green, M. D. (1990). Neandertal craniofacial growth:
an ontogenetic model. Unpublished M.A. thesis,
University of Tennessee, Knoxville.
Green, M. D. & Smith, F. H. (1991). Heterochrony
and Neandertal craniofacial morphogenesis. Am. J.
phys. Anthrop. Supplement 12, 12 (abstract).
Guglielmino-Matessi, C., Gluckman, P. & Cavalli-
Sforza, L. (1979). Climate and the evolution of skull
metrics in man. Am. J. phys. Anthrop. 50, 549–564.
Hall, B. K. (1978). Developmental and Cellular Skeletal
Biology. New York: Academic Press.
Herring, S. W. (1993). Epigenetic and functional influences
on skull growth. In (J. Hanken & B. K. Hall,
Eds) The Skull, vol. 1, pp. 153–206. Chicago:
University of Chicago Press.
Hooton, E. A. (1946). Up from the Ape. New York:
Howells, W. W. (1969). Criteria for selection of Osteometric
dimensions. Am. J. phys. Anthrop. 30, 451–
Howells, W. W. (1973). Cranial Variation in Man.
Cambridge: Peabody Museum Papers, No. 67.
Hylander, W. L. (1988). Implications of in vivo experiments
for interpreting the functional significance of
      313
‘‘Robust’’ australopithecine jaws. In (F. L. Grine,
Ed.) Evolutionary History of the Robust Australopithecines,
pp. 55–80. Chicago: Aldine.
Hylander, W. L. & Johnson, K. R. (1992). Strain
gradients in the craniofacial region of primates. In
(Z. Davidovitch, Ed.) The Biological Mechanisms of
Tooth Movement, pp. 559–569. Columbus, OH: Ohio
State University College of Dentistry.
Jacobsen, A. (1993). Somitomeres: Mesodermal segments
of the head and trunk. In (J. Hanken & B. K.
Hall, Eds) The Skull, vol. 1, pp. 42–111. Chicago:
University of Chicago Press.
Jelı´nek, J. (1969). Neanderthal man and Homo sapiens
in central and eastern Europe. Curr. Anthrop. 10,
Kiliaridis, S., Engstro¨m, C. & Thilander, B. (1986).
The relationship between masticatory muscle function
and cranial morphology. I. A cephalometric
longitudinal analysis in the growing rat fed a soft diet.
Eur. J. Orthod. 7, 273–283.
Kohn, L. A. P., Leigh, S. R., Jacobs, S. C. & Cheverud,
J. M. (1993). Effects of annular cranial vault modification
on the cranial base and face. Am. J. phys.
Anthrop. 90, 147–168.
Kohn, L. A. P., Leigh, S. R. & Cheverud, J. M. (1995).
Asymmetric vault modification in Hopi crania. Am.
J. phys. Anthrop. 98, 173–196.
Le Douarin, N. M., Ziller, C. & Couly, G. F. (1993).
Patterning of neural crest derivatives in the avian
embryo: in vivo and in vitro studies. Develop. Biol.
159, 24–49.
Lieberman, D. E. (1995). Testing hypotheses about
recent human evolution from skulls: integrating
morphology, function, development and phylogeny.
Curr. Anthrop. 36, 159–197.
Lieberman, D. E. (1996). How and why recent humans
grow thin skulls: experimental data on systemic cortical
robusticity. Am. J. phys. Anthrop. 101, 217–236.
Lieberman, D. E. (1998). Sphenoid shortening and the
evolution of modern human cranial shape. Nature
393, 158–162.
Lieberman, D. E. (1999). Homology and hominid
phylogeny: problems and potential solutions. Evol.
Anthrop. 7, 142–151.
Lieberman, D. E., Pilbeam, D. R. & Wood, B. A.
(1996). Homoplasy and early Homo: an analysis of
the evolutionary relationships of H. habilis sensu stricto
and H. rudolfensis. J. hum Evol. 30, 97–120.
Lilliefors, H. W. (1967). On the Kolmogorov–Smirnov
test for normality with means and variance unknown.
J. Am. Stat. Assoc. 64, 399–402.
Linder-Aronson, S. (1979). Naso-respiratory function
and craniofacial growth. In (J. A. McNamara Jr.,
Ed.) Naso-Respiratory Function and Craniofacial
Growth, pp. 121–147. Ann Arbor: Center for Human
Growth and Development, Monograph No. 9.
Martin, R. & Saller, K. (1956). Lehrbuch der Anthropologie.
Stuttgart: Gustav Fischer Verlag.
McNeill, R. W. & Newton, G. M. (1965). Cranial
base morphology in association with intentional
vault deformation. Am. J. phys. Anthrop. 23, 241–
Moore, W. J. & Lavelle, C. L. B. (1974). Growth of the
Facial Skeleton in the Hominoidea. London: Academic
Moss, M. L. (1958). The pathogenesis of artificial
cranial vault deformation. Am. J. phys. Anthrop. 16,
Moss, M. L. & Young, R. W. (1960). A functional
approach to craniology. Am. J. phys. Anthrop. 18,
Moyers, R. E. (1988). Handbook of Orthodontics, 4th
edn. Chicago: Year Book Medical Publishers.
Olson, T. R. (1981). Basicranial morphology of the
extant hominoids and Pliocene hominids: The new
material from the Hadar Formation, Ethiopia, and its
significance in early human evolution and taxonomy.
In (C. B. Stringer, Ed.) Aspects of Human Evolution,
pp. 99–128. London: Taylor & Francis.
Plourde, A. M. & Anto´n, S. C. (1992). Craniofacial
compensation in artificially deformed children’s
crania. Am. J. phys. Anthrop. (Suppl. 14) 132.
Principato, J. & Wolff, P. (1985). Pediatric nasal air
resistance. Laryngoscope 95, 1067–1069.
Ranly, D. M. (1980). A Synopsis of Craniofacial Growth.
New York: Appleton-Century-Crofts.
Ravosa, M. J. (1991). Interspecific perspective on
mechanical and nonmechanical models of primate
circumorbital morphology. Am. J. phys. Anthrop. 86,
Richards, G. D. & Anto´n, S. C. (1991). Craniofacial
configuration and postcranial development of a
hydrocephalic child (ca. 2500 B.C.–500 A.D.) with a
review of cases and comments on diagnostic criteria.
Am. J. phys. Anthrop. 85, 185–200.
Ross, C. F. & Ravosa, M. J. (1993). Basicranial flexion,
relative brain size, and facial kyphosis in nonhuman
primates. Am. J. phys. Anthrop. 91, 305–324.
Ross, C. & Henneberg, M. (1995). Basicranial flexion,
relative brain size, and facial kyphosis in Homo sapiens
and some fossil hominids. Am. J. phys. Anthrop. 98,
Scott, J. H. (1958). The cranial base. Am. J. phys.
Anthrop. 16, 319–348.
Selleck, M. A. J., Scherson, T. Y. & Bronner-Fraser,
M. (1993). Origins of neural crest cell diversity.
Develop. Biol. 159, 1–11.
Shaffer, H. B., Clark, J. M. & Kraus, F. (1991). When
molecules and morphology clash: a phylogenetic
analysis of the North American Ambystomatid salamanders
(Caudata: Ambystomatidae). Syst. Zool. 40,
Shea, B. T. (1985). On aspects of skull form in African
apes and orangutans, with implications for hominid
evolution. Am. J. phys. Anthrop. 68, 329–342.
Sirianni, J. E. (1985). Nonhuman primates as models
for human craniofacial growth. In (E. S. Watts, Ed.)
Nonhuman Primate Models for Human Growth and
Development, pp. 95–124. New York: Alan R. Liss.
Sirianni, J. E. & Swindler, D. R. (1979). A review of
postnatal craniofacial growth in old world monkeys
and apes. Yearb. Phys. Anthrop. 22, 80–104.
Sjøvold, T. (1984). A report on the heritability of
some cranial measurements and non-metric traits. In
314 . .  ET AL.
(G. N. Van Vark &W.W. Howells, Eds) Multivariate
Statistics in Physical Anthropology, pp. 223–246.
Dordrecht: D. Reidel.
Skelton, R. R. & McHenry, H. M. (1992). Evolutionary
relationships among early hominids. J. hum. Evol.
23, 309–349.
Smith, F. H. (1982). Upper Pleistocene hominid evolution
in South-Central Europe: a review of the
evidence and analysis of trends. Curr. Anthrop. 23,
Smith, F. H. (1984). Fossil hominids from the Upper
Pleistocene of Central Europe and the origin of
modern Europeans. In (F. H. Smith & F. Spencer,
Eds) The Origins of Modern Humans: A World Survey
of the Fossil Evidence, pp. 137–209. New York: Alan
R. Liss.
Smith, F. H. (1994). Samples, species and speculations
in the study of modern human origins. In (M. H.
Nitecki & D. V. Nitecki, Eds) Origins of Anatomically
Modern Humans, pp. 227–249. New York:
Smith, F. H. & Green, M. D. (1991). Heterochrony,
life history, and Neandertal morphology. Am. J. phys.
Anthrop. Supplement 12, 164 (abstract).
Sokal, R. F. & Rohlf, F. H. (1995). Biometry, 3rd edn.
New York: W. H. Freeman.
Sokal, R. R. & Sneath, P. H. A. (1963). Principles
of Numerical Taxonomy. San Francisco: W. H.
Sperber, G. H. (1989). Craniofacial Embryology, 4th
edn. London: Wright.
Strait, D. S. (1998). Evolutionary integration in the
hominid cranial base. Ph.D. Dissertation, State
University of New York, Stony Brook.
Strait, D. S., Grine, F. E. & Moniz, M. A. (1997). A
reappraisal of hominid phylogeny. J. hum. Evol. 32,
Stramrud, L. (1959). External and internal cranial
base: A cross-sectional study of growth and of association
with form. Acta Odont. Scand. 17, 239–266.
Thorogood, P. (1993). Differentiation and morphogenesis.
In (J. Hanken & B. K. Hall, Eds) The Skull,
vol. 1, pp. 112–152. Chicago: University of Chicago
Trinkaus, E. & LeMay, M. (1982). Occipital bunning
in Late Pleistocene hominids. Am. J. phys. Anthrop.
57, 27–35.
Vandermeersch, B. (1981). Les Hommes Fossiles de
Qafzeh (Isra¨el). Paris: C.N.R.S.
Vlcˇek, E. (1991). L’homme fossile en Europe Centrale.
L’Anthropol. 95, 409–472.
Warren, D. W., Hairfield, W. M., Seaton, D. L. &
Hinton, V. A. (1987). The relationship between
nasal airway cross-sectional area and nasal resistance.
Am. J. Dentofac. Orthop. 92, 390–395.
Watts, E. S. (1985). Adolescent growth and development
of monkeys, apes and humans. In (E. S. Watts,
Ed.) Nonhuman Primate Models for Human Growth
and Development, pp. 41–65. New York: Alan R. Liss.
Weidenreich, F. (1941). The brain and its role in the
phylogenetic transformation of the human skull.
Trans. Am. Phil. Soc., N.S. 31, 321–442.
Weidenreich, F. (1945). The brachycephalization of
recent mankind. Southwest. J. Anthrop. 1, 1–54.
Wolpoff, M. (1996). Human Evolution. New York:
McGraw Hill.
Wood, B. A. (1991). Koobi Fora Research Project, Volume
4: Hominid Cranial Remains. Oxford: Oxford
University Press.
      315

Saturday, October 23rd, 2004, 11:35 PM
I have to question their conclusion that bunning in early moderns is distinct from that in neanderthals because apart from one of them they have a bun index of 4, compared to the neanderthal bun index of 5, and therefore their bunning does seem close to that of neanderthals, more than it seems to follow the pattern in the later moderns.