Integration and Homology of “Chignon” and “Hemibun” Morphology

Chapter 17
Integration and Homology of “Chignon” and “Hemibun” Morphology
Philipp Gunz and Katerina Harvati
Abstract The presence of a weak occipital bun in some
Upper Paleolithic European fossils is often cited as evidence
for admixture between Neanderthals and anatomically modern humans, because the “chignon” morphology is considered by many to be a derived Neanderthal trait.
It is impossible, however, to split this morphology into
“present” or “absent” character states (and thus “primitive”
or “derived”); it rather varies in continuous degrees of
expression. Furthermore the shape of the upper scale of the
occipital bone is tightly integrated with the shape of the other
bones forming the vault. To assess whether the “hemibun” of
some Upper Paleolithic European crania should be considered evidence for possible hybridization, it is thus crucial to
understand the integration of this morphology and whether
this shape feature is homologous between modern humans
and Neanderthals. Here we present a geometric morphometric analysis assessing the integration of the posterior midsagittal profile and the temporal bone quantitatively. We digitized
3-D coordinates of anatomical landmarks on the posterior
vault and semilandmarks along a midsagittal curve from
bregma to inion on 356 modern and archaic human crania.
These points were converted into shape coordinates using
Procrustes superimposition and then analyzed using the
method of singular warps. The occurrence of an occipital bun
is highly correlated with a flat parietal midline and an anteriorly positioned temporal bone. While Upper Paleolithic
Homo sapiens cannot be distinguished from recent humans,
archaic Homo fall outside the range of modern variation. The
pattern of integration however, which accounts for ~30% of
the total variation, is shared between modern humans and
archaic Homo. Our results suggest that the occurrence of
P. Gunz (*)
Department of Human Evolution, Max Planck Institute for Evolutionary
Anthropology, Deutscher Platz 6, D-04103 Leipzig, Germany
e-mail: gunz@eva.mpg.de
K. Harvati
Department of Early Prehistory and Quaternary Ecology,
Senckenberg Center for Human Evolution and Paleoecology,
Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070
Tübingen, Germany
e-mail: katerina.harvati@ifu.uni-tuebingen.de
“hemibuns” in UPE should not be used as evidence for
admixture between modern humans and Neanderthals, as
this morphology is a predictable correlate of the relative
position of the temporal bone and not an independent trait.
Keywords Procrustes analysis • Homo • Hybridization
• Singular warps • Partial Least Squares • Homo
neanderthalensis
Introduction
The “chignon” or “occipital bun” morphology, defined as
a posterior projection or a great convexity of the upper
scale of the occipital bone, is one of the features often cited
as ­evidence for admixture between modern humans and
Neanderthals. Because the “chignon” is considered by many
to be a derived Neanderthal trait (Hublin 1988; Dean et al.
1998; see also Lieberman 1995) and because some Upper
Palaeolithic crania from Europe exhibit a similar, yet less pronounced, morphology (sometimes called a “hemibun”) some
authors have argued that this indicates interbreeding between
Neanderthals and early modern Europeans (Jelinek 1969;
Genet-Varcin 1970; Vlcek 1970; Smith 1982, 1984; Bräuer
1989; Gambier 1997; Churchill and Smith 2000; Wolpoff
et al. 2001; Smith et al. 2005). However, since the “chignon”
is a morphological continuum, it is difficult to split into “present” or “absent” character states (Trinkaus and LeMay 1982),
and therefore to be considered a trait in a cladistic sense. The
putative evidence for interbreeding is thus not the presence of
a derived Neanderthal discrete trait in Upper Paleolithic modern Europeans; instead, it is the simple assumption that a
hybrid of two morphologically distinct populations, one with
a distinct “chignon”, the other one with a more globular cranium, would show an average morphology.
Using a geometric morphometric analysis of points measured along the midline of the posterior profile, we recently
studied the chignon quantitatively and assessed its usefulness
in separating Neanderthals from modern humans (Gunz and
Harvati 2007; see also Harvati 2001; Harvati et al. 2002;
Reddy et al. 2005). In that study we showed that it is not
S. Condemi and G.-C. Weniger (eds.), Continuity and Discontinuity in the Peopling of Europe: One Hundred Fifty Years of Neanderthal Study, 193
Vertebrate Paleobiology and Paleoanthropology, DOI 10.1007/978-94-007-0492-3_17, © Springer Science+Business Media B.V. 2011
194
p­ ossible to metrically separate modern humans and archaic
Homo, including Neanderthals, based on the midline shape of
the upper scale of the occipital bone alone. It is rather the
position and orientation of the occipital bone relative to the
parietal and temporal bones, or to put it differently, the overall
shape of the posterior half of the neurocranium, that constitutes the unique Neanderthal morphology. Our results concurred with previous work that this condition is the result of
an integrated shape change and that differences in the shape
of the posterior profile are probably related to differences in
the rate of brain growth and timing of suture closure (Trinkaus
and LeMay 1982; Lieberman 1995; Lieberman et al. 2000).
Our findings, however, differed from those of Lieberman
et al. (2000), who observed a correlation between occipital
bunning and narrow cranial bases in modern humans, and the
inverse relationship in Neanderthals. Lieberman et al. (2000)
tentatively concluded that bunning in Neanderthals and Upper
Paleolithic Europeans is not homologous, and hence its presence in the latter does not indicate genetic continuity. Contrary
to Lieberman et al. (2000), we (Gunz and Harvati 2007) failed
to detect differences in integration between modern and archaic
humans, and concluded that the ­“chignon” in Neanderthals
and “hemibun” morphology in modern humans were homologous. In this chapter we revisit our previous analysis and elaborate on methodological details that received little attention in
the original publication. Furthermore we discuss our method
and results in light of the recent work on modularity and integration by Mitteroecker and Bookstein (2007).
Aim of This Study
Our goal was to reassess the hypothesis that Neanderthals
and modern humans exhibit different patterns of cranial integration resulting in non-homologous posterior cranial projection. For this purpose we evaluated the influence of the
shape of the posterior midsagittal profile and the cranial base
on the expression of occipital bunning in our combined
recent and fossil human sample. We quantified how tightly
the bones forming the posterior vault are integrated and
whether Neanderthals and modern humans follow different
trajectories, or form distinct clusters along the same trajectory. We then repeated the analysis looking at the recent
human and fossil samples separately.
In order to quantify the shape of the posterior vault we
used geometric morphometrics based on coordinate data,
measured with a 3-D digitizer on both temporal bones
and the midsagittal profile. Using homologous points
(semilandmarks) along the midsagittal curve enabled us to
quantify the continuous variation in the expression of the
midsagittal aspect of the “chignon” morphology. We explored
P. Gunz and K. Harvati
the co-variation of the shape of the posterior vault and the
cranial base (as represented by our parietal and temporal
bone landmarks and semilandmarks) on the expression of
occipital bunning using the method of singular warps
(Bookstein et al. 2003). If modern humans and Neanderthals
differed in their pattern of cranial integration, and hence did
not share homologous occipital buns, as suggested by
Lieberman et al. (2000), the two groups would not follow the
same predictions of occipital shape based on the shape and
relative positions of the parietal and temporal bones.
In this chapter we specifically address two questions that
were not explored in full detail in our previous study of the
same data-set: (1) whether the small sample of fossils biases
the results, and (2) whether partitioning the midsagittal curve
into two blocks exaggerates the correlation between these
two blocks.
Because the number of extant modern humans is more
than ten times the number of fossil specimens, one could easily imagine situations where the analysis of the pooled sample would impose the modern integrational pattern upon the
fossils, even if the modern and fossil patterns were distinct.
To check whether the excess of modern specimens would
bias the results, we performed a separate analysis of all fossil
specimens.
The second concern is slightly more technical: The midsagittal profile is a single curve, not two independent curves.
Partioning it into two parts at an arbitrary point to look for
covariation (as we did in Gunz and Harvati 2007), will naturally yield high correlations between these two blocks.
Furthermore the semilandmarks along the posterior profile
were allowed to slide along the curve so as to minimize the
thin-plate spline bending energy before partitioning into
blocks. Because these semilandmarks slide along the same
curve, this sliding step potentially exaggerates correlations
among them. In the present study we have tried to avoid
those potential pitfalls by treating the midsagittal profile as
one curve, thereby reducing the analysis from a three-block
to a two-block design.
Materials and Methods
Samples
This study included a large comparative sample of recent
humans (n = 326) and Middle and Late Pleistocene fossil
specimens (Table 17.1). Recent humans were represented by
ten regional groups from a wide geographical range, including an Iberomaurusian series from Afalou and Taforalt (North
Africa), dated to 14–8.5 ka (Chamla 1978). Only adult crania
with fully erupted permanent dentition were measured. Most
17 Integration of the Neanderthal “Chignon”
195
Table 17.1 Sample of recent humans and fossil specimens used in this
analysis
Fossil humans
(n = 29)
Middle-Late Pleistocene African and near
Eastern fossil humans (MLA)
Kabwe
Ndutu
Ngaloba
Omo 2
Qafzeh 9
Skhul 5
Singa
(n = 7)
Middle Pleistocene European
Pre-Neanderthals (MPE)
Reilingen
Sima de los Huesos 5*
(n = 2)
‘Classic’ and Early Neanderthals (NE)
Amud 1
Guattari 1
La Chapelle-aux-Saints
La Ferrassie 1
La Quina 5
Saccopastore 1
Shanidar 1*
Spy 1 and Spy 10
Tabun
(n = 10)
Eurasian Upper Paleolithic specimens (UP)
Abri Pataud
Chancelade*
Cioclovina
Cro Magnon 1,2
Ein Gev
Mladeč 1 and 5
Předmostí 3* and 4*
(n = 10)
Recent humans
(n = 326)
1. African
(n = 36)
(Mali, Kenya)
2. Andaman
(n = 30)
(Andaman Islands)
3. Asian
(n = 38)
(North China, Thailand)
4. Australian
(n = 32)
(South Australia)
5. Inuit
(n = 42)
(Alaska, Greenland)
6. European
(n = 45)
(Austria, Greece, Italy, Germany, Yugoslavia)
(n = 28)
7. Khoi-San
(South Africa)
(n = 30)
8. Melanesian
(New Britain)
(n = 19)
9. Middle Eastern
(Syria)
(n = 26)
10.Iberomaurusian
(Afalou, Taforalt)
* Asterisks indicate specimens for which casts or stereolithographs
were used
Table 17.2 Landmarks collected on the temporal, parietal and occipital
bones
Temporal block (L + R)
Number of
landmarks: 18
Porion
Auriculare
Mastoidiale
Stylomastoid Foramen
Most postero-lateral point of the jugular
fossa
Lateral origin of the petro-tympanic crest
Most medial point of the petro-tympanic
crest at the level of the carotid canal
Most inferior point on the juxtamastoid
crest (following Hublin 1978)
Deepest point of the lateral margin of the
articular eminence
(root of the articular eminence)
Midsagittal Block
Number of
landmarks and
semilandmarks: 37
Bregma
20 semilandmarks between
Bregma and Lambda
Lambda
14 semilandmarks between
Lambda and Inion
Inion
See also Fig. 17.1
crania were only sexed based on anthropological criteria, so
the sexes were pooled in the analysis. The fossil sample comprised ten Neanderthal specimens from Europe and the Near
East; two European Middle Pleistocene specimens; seven
Middle-Late Pleistocene fossils from Africa and the Near
East; and ten Upper Paleolithic anatomically modern humans
from Europe and the Near East (Table 17.2). In the few cases
where the original specimens were not available, high quality
casts or stereolithographs from the Anthropology Depart­
ments of the American Museum of Natural History and of
New York University were measured.
Data
3-D coordinates of anatomical landmarks on the temporal,
parietal and occipital bones as well as semilandmarks along
the midsagittal from bregma to inion were collected by KH
using a Microscribe 3DX digitizer. Each cranium was measured in a ‘top’ and ‘bottom’ orientation, which were later
superimposed using four fiducial points, which were not used
in the analysis. The midsagittal profile was digitized as closely
spaced points along the curve. These points were then automatically resampled to yield equal point count on every specimen: we computed a cubic spline through the points measured
196
P. Gunz and K. Harvati
Fig. 17.1 Landmarks and semilandmarks used in this analysis. For the singular warps analysis the set of landmarks and semilandmarks was
­partitioned into two blocks: (1) the midsagittal profile from bregma to lambda to inion, (2) the landmarks on the temporal bones (both sides)
on the midsagittal curve and then placed 20 equidistantly
spaced points between bregma and lambda, and 14 between
lambda and inion. Since morphometric analyses do not accommodate missing data, some data reconstruction was necessary
for the incomplete fossils. During data collection, and only for
specimens with minimal damage, landmarks were reconstructed using anatomical information from the preserved surrounding areas. In cases where bilateral landmarks were
missing on one side only, they were estimated by reflected
relabeling (Mardia and Bookstein 2000; see Gunz 2005; Gunz
et al. 2009 regarding application to missing data).
Incomplete specimens were least-squares superimposed
with their reflected configurations in Procrustes space and
missing data were reconstructed from their homologous
counterparts on the other side. Landmarks on those specimens
that had only one side of temporal bone preserved, or were
one side was affected by distortion were reflected by mirroring
across the empirical midplane (Reilingen, Předmostí 4,
Kabwe, Omo2, LH18, Cro Magnon2, Spy10, Singa).
In limited instances (specimens Cro Magnon2, Ndutu,
Omo2, Mladeč1, Saccopastore1), points missing on both sides
of the cranium or on the midsagittal plane were estimated by
minimizing the bending energy of the thin-plate spline
between the incomplete specimen and the sample Procrustes
average, following (Gunz 2005; Gunz et al. 2009).
The semilandmarks were iteratively allowed to slide along
the midsagittal curve to minimize the bending energy of the
thin-plate spline interpolation function computed between
each specimen and the sample Procrustes average. We used
the algorithm of Bookstein (1997; see also Gunz et al. 2005,
2009), which allows points to slide along tangents to the curve.
These tangents were approximated for each semilandmark by
converting the vector between the two neighboring points to
unit length. Missing points were allowed to slide without
­constraining them to a curve (“full relaxation”). Splinerelaxation removes the effects of “digitizing error” in the tangent direction that results from the practical necessity of
having to place the semilandmarks somewhere along the
curve. After relaxation these semilandmarks can be treated in
multivariate analyses as if they had been homologous points
in the first place (Bookstein 1997; Bookstein et al. 1999; Gunz
et al. 2005). Landmarks and slid semilandmarks were converted to shape coordinates by Procrustes superimposition
(GPA: Gower 1975; Rohlf and Slice 1990). This procedure
removes information about location and orientation from the
raw coordinates and standardizes each specimen to unit centroid size, a size-measure computed as the square root of the
summed squared Euclidean distances from each landmark to
the specimen’s centroid (Dryden and Mardia 1998).
Unlike in Gunz and Harvati (2007) the landmark and
semilandmark data were partitioned into two blocks rather
than three: temporal and midsagittal (Table 17.2, Fig. 17.1).
The “midsagittal-block” included the landmarks and
semilandmarks defining the posterior midline cranial profile
from bregma to lambda to inion. The “temporal bone block”
included landmarks on the basicranial portion of both right
and left temporal bones (see Table 17.2). The temporal landmarks were measured on both sides in order to include information about the width of the cranial base.
Analyses
We studied the morphological integration of the cranial
base and the posterior midline profile using the method of
­“singular warps” introduced by Bookstein et al. (2003) that
197
17 Integration of the Neanderthal “Chignon”
quantifies and visualizes the covariation of anatomical
regions. Blocks of landmarks were defined a priori; then the
linear combinations of the original shape variables that
provide the best mutual cross-prediction between these
landmark-blocks were calculated. Usually a singular warps
analysis with two blocks is calculated via a singular value
decomposition of the cross-block covariance matrix
(Bookstein et al. 2003). Here we used a more flexible algorithm suggested by Mitteroecker and Bookstein (2007),
based on a “Wright-style” factor analysis (Wright 1932).
The results are identical to the SVD approach, when the singular vectors are scaled appropriately. We will return to this
numerical identity between these two approaches in the
­discussion section. Mitteroecker and Bookstein (2007) demonstrate that the singular vectors have to be scaled before
joint visualisation. This scaling was performed in the ­analysis
of Gunz and Harvati (2007) but not described in the paper.
Higher singular warps were calculated using the same
­algorithm after projecting out the lower singular vectors for
each block separately.
Singular warps are linear combinations of the original
shape variables and can be visualized either as scores, or as
deformations. We scaled the singular vectors following
Mitteroecker and Bookstein (2007) and visualized the
deformations of the temporal and midsagittal block
together. To create the surface-morphs the landmarks and
semilandmarks of one modern human specimen were used
to warp the triangulated surface from the mean configuration in Procrustes space to this mean shape with different
multiples of the singular vectors added. Note that the surface areas where there is no (semi)landmark information
are just smoothly warped according to the thin-plate spline
interpolation.
For data processing and analyses we used software routines written in Mathematica (© Wolfram Research) and “R”
by PG together with Philipp Mitteroecker (University of
Vienna).
Results
Singular Warps: Pooled Sample
The scores of the first singular warp (which explains ~30%
of the total variation) are plotted in Fig. 17.2 (scores for midsagittal and temporal blocks together as well as separately
for each group); the associated shape changes are visualized
in Fig. 17.3. The integration of the midsagittal profile and the
position and shape of the temporal bones is very tight (the
correlation coefficient of the midsagittal and temporal scores
r > 0.89); the points lie along a single trajectory. The fossil
specimens do not diverge from this common pattern of
­integration. Looking at the scores for each block and “population” separately shows that the variation along these axes
within each group is considerable. Neanderthals and the
other fossil human groups have on average higher scores
than the modern mean, but almost all modern groups have
specimens that fall well within the Neanderthal range of
variation.
Figure 17.3 visualizes the shape change associated with
the first singular warp as a surface-morph. The singular vectors are scaled following Mitteroecker and Bookstein (2007)
so that they can be visualized jointly. An anteriorly and
superiorly placed cranial base predicts (and is predicted by)
a flat parietal and a posterior projection of the occipital.
Along the first singular warp, there are only subtle shape
changes of the temporal bone; it is the relative position of
the cranial base that changes. The higher singular warps
were calculated after projecting out the lower singular vectors for every block separately. The plots are not shown here,
as the associated shape changes do not involve occipital
bunning.
Because the sample comprises so many more modern
humans than fossil specimens, we checked carefully whether
this shared trajectory could be an artifact due to the unequal
sample sizes. There exists another potential pitfall with
pooled samples: if populations had different means, then
even if there were different patterns of integration among
populations, looking at integration based on the uncentered
pooled covariance matrix would create the impression of a
common trajectory. To avoid this, we subtracted the respective group mean from each specimen before computing the
covariance matrix for the singular warps.
Singular Warps: Fossil Sample
Figure 17.4 shows the scores of the first and second singular
warp of the fossil sample. In essence the results stay the same
as those for the pooled analysis; in the fossil sample the
shape-changes described above for the pooled analysis are
captured by the first two singular warps: the first singular
warp is driven by the correlation of a flat parietal midline
with a relatively wide cranial base (r ~ 0.82); the second singular warp entails posterior projection of the occipital midline, an anteriorly and superiorly positioned temporal bone
and a relatively anterior placement of bregma (correlation
r > 0.62). The scores of the first singular warp (upper left
panel) show two clusters along the midsagittal scores. Only
the second singular warp is associated with occipital bunning, and the corresponding scores (upper right panel) show
no differences between anatomically modern humans and
archaic Homo in these dimensions.
198
P. Gunz and K. Harvati
Fig. 17.2 Pooled sample; scores of the first singular warp of ­midsagittal
and temporal block plotted against each other. These plots show how
well one can predict the shape of one block from the other block. NE
and MPE crania have higher scores than most modern humans for all
blocks, but the integrational pattern is the same between groups. The
scores are plotted for each group separately to demonstrate the considerable variability within geographic groups. The numbers correspond to
the order of modern human populations in Table 17.1
Shape Regression
width of the cranial base as the length of the vector between
left and right porion of the original coordinates before superimposition, then calculated the multivariate regression of all
Procrustes shape coordinates on this width. When visualized
by adding the regression slopes to a mean shape (Fig. 17.5),
an absolute increase in cranial base width in recent humans
predicts a projection of the occipital bone and a slightly flatter parietal. In the same way we predicted all shape variables
Because Lieberman et al. (2000) found occipital bunning in
recent humans to be correlated with narrow cranial bases, we
also looked at shape predictions for modern humans with
absolutely narrower cranial bases (in contrast to the relative
shape differences in Procrustes space discussed above). For
all extant Homo sapiens crania we computed the absolute
17 Integration of the Neanderthal “Chignon”
199
Fig. 17.3 Pooled sample; first singular warp. The shape dimensions
associated with the scores shown in Fig. 17.2 can be visualized together.
Here a modern human cranium’s surface is deformed by subtracting
(left) and adding (right) a multiple of the singular vectors to the mean
shape (middle). The first singular warp contrasts a globular cranium
with a more elliptical shape. Note that as occipital bunning increases,
the cranial width does not change at all; the temporal bones get more
superiorly and anteriorly placed and the mastoid decreases in size
by the logarithm of centroid size, as a crude proxy for
endocranial volume. An increase in absolute size predicts a
slightly more globular shape of the posterior vault and an
occipital protuberance, but no occipital bun.
There exists considerable shape variability within groups
of modern humans, but the pattern of integration is shared
among modern human populations, fossil modern human
specimens and archaic Homo. While Neanderthals and
archaic Homo consistently had higher scores than most modern humans, they followed the same linear trend when the
two blocks were viewed together. Neanderthals thus have the
amount of occipital bunning that one would predict for a
human with such a supero-anteriorly positioned temporal
bone and such a flat parietal. This is particularly apparent in
the modern humans that have scores similar to archaic
humans. The Upper Paleolithic specimens all plot within the
modern human cloud. Note that – with the exception of the
Andamenese (population #2) – every modern human group
has outliers that fall within the Neanderthal variation along
the midsagittal scores of the first singular warp (Fig. 17.2).
This includes crania from Africa, Asia and the Middle East,
hence no uniquely European pattern can be discerned.
Changing the singular warp analysis from the three-block
design of Gunz and Harvati (2007) to a two-block design had
no impact on the results. Analyzing the fossil specimens separately confirmed the conclusions of our original publication
that the fossils and extant specimens follow the same shape
prediction between the relative position of the temporal bones
and the occipital profile. A posterior projection of the occipital bone is associated with a relatively anterior and superior
position of the temporal bone. We would like to stress however, that this conclusion only pertains to the external aspects
Discussion
The results of our singular warps analysis confirm our previous findings (Gunz and Harvati 2007), indicating that occipital bunning in modern humans and Neanderthals alike is
associated with anteriorly and superiorly placed temporal
bones/cranial bases (Figs. 17.3 and 17.4). The shape of the
posterior midline profile and the relative position of the cranial base are tightly integrated; the scores of the two landmark blocks are highly correlated and fall along a common
trajectory in those dimensions related to occipital bunning.
These high correlations are unlikely to be caused by direct
individual interactions of the bones; instead they are most
likely spurious correlations because all bones are affected by
a single factor, brain expansion, during ontogeny. Mitteroecker
(2007) and Mitteroecker and Bookstein (2007) show that the
results of a two-block partial least squares analysis and
Wrigth (1932) factor analysis are identical (when scaled
appropriately). Thus the results of the present analysis can be
interpreted as loadings of the two landmark modules on a
common factor.
Fig. 17.4 Fossil sample; first and second singular warp. Left: Scores of
the first singular warp computed from the fossil specimens only. There
are two clusters along the midsagittal scores along a general trend. The
bottom panel shows the associated shape change as vectors: increased
width of the cranial base co-varies with a flat parietal profile. Note that
there is almost no shape change in the occipital profile. Right: One
common trajectory with complete overlap of modern and archaic
humans. An projecting occipital bun is predicted by an anterior and
superior placement of the cranial base. Figure 17.4b visualizes the same
shape difference as a surface morph
201
17 Integration of the Neanderthal “Chignon”
Fig. 17.5 Shape predictions for a wide cranial base in modern humans.
Overall shape was regressed on the absolute length of the vector
between left and right porion. An absolutely wider cranial base predicts
a posterior projection of the midsagittal profile in extant modern
humans. The regression slopes are visualized as vectors (left) from
the mean shape towards an increase in cranial base width. Mean shape
(middle) and positive extreme shape (right) are also shown as a surface
morph
of the occipital bun in the midsagittal plane, as there were no
endocranial landmarks used in this study.
stages by grants to KH by the American Museum of Natural History;
NYCEP; the Onassis and the CARE Foundations; and the U.S. National
Science Foundation. Support was also provided by New York University,
the Max Planck Institute for Evolutionary Anthropology and the
“EVAN” Marie Curie Research Training Network MRTN-CT-019564.
This is NYCEP morphometrics contribution number 34.
Conclusions
We could not detect differences in integration between
­modern and fossil humans with respect to the external aspects
of midsagittal posterior projection and relative width and
position of the temporal bones. We thus consider the midsagittal aspect of the “chignon” morphology in Neanderthals to
be homologous to bunning morphology in modern humans.
The occurance of “hemibuns” in Upper Paleolithic modern
Europeans should not be used as evidence for admixture
between modern humans and Neanderthals, as this morphology is a predictable correlate of the relative position of the
cranial base and not an independent trait.
Acknowledgments We thank all the curators and collections managers
in several institutions across Europe, Africa and the USA, for kindly
allowing access to both fossil and extant material used in this study.
We thank Maximilian v. Harling for the CT scan used to create surface
morphs, and Silvana Condemi for her efforts putting together this
volume. We are also grateful to Jean-Jacques Hublin, Tim Weaver, Fred
Bookstein, Philipp Mitteröcker, Markus Bastir, Susan Antón, Dan
Lieberman and two anonymous reviewers for providing very helpful
comments and suggestions. This research was funded in its various
References
Bookstein, F. L. (1997). Landmark methods for forms without
­landmarks: Morphometrics of group differences in outline shape.
Medical Image Analysis, 1(3), 225–243.
Bookstein, F. L., Schaefer, K., Prossinger, H., Seidler, H., Fiedler, M.,
Stringer, C. B., Weber, G. W., Arsuaga, J. L., Slice, D., Rohlf, F. J.,
et al. (1999). Comparing frontal cranial profiles in archaic and modern Homo by morphometric analysis. Anatomical Record, 257(6),
217–224.
Bookstein, F. L., Gunz, P., Mitteroecker, P., Prossinger, H., Schaefer,
K., & Seidler, H. (2003). Cranial integration in Homo: Singular
warps analysis of the midsagittal plane in ontogeny and evolution.
Journal of Human Evolution, 44(2), 167–187.
Bräuer, G. (1989). The evolution of modern humans: A comparison
between the African and non-African Evidence. In C. B. Stringer &
P. Mellars (Eds.), The human revolution (pp. 123–154). Princeton:
Princeton University Press.
Chamla, M. C. (1978). Le peuplement de l’Afrique duNord de
l’Epipaleolithique a l’ epoque actuelle. L’Anthropologie, 82, 385–430.
Churchill, S. E., & Smith, F. H. (2000). Makers of the Early Aurignacien
of Europe. American Journal of Physical Anthropology (Yrbk), 43,
61–115.
202
Dean, D., Hublin, J. J., Holloway, R., & Ziegler, R. (1998). On the
­phylogenetic position of the pre-Neandertal specimen from
Reilingen, Germany. Journal of Human Evolution, 34(5), 485–508.
Dryden, I., & Mardia, K. V. (1998). Statistical shape analysis.
New York: Wiley.
Gambier, D. (1997). Modern humans at the beginning of the
Upper Paleolithic in France; anthropological data and perspectives.
In G. A. Clark & C. M. Willermet (Eds.), Conceptual issues in
­modern human origins research (pp. 117–131). New York: Aldine
de Gruyter.
Genet-Varcin, E. (1970). Considérations morphologiques sur l’homme
de Cro-Magnon. In G. Camps & G. Olivier (Eds.), l’homme de
­Cro-Magnon. Paris: Arts et Métiers Graphiques.
Gower, J. C. (1975). Generalized procrustes analysis. Psychometrika,
40, 33–51.
Gunz, P. (2005). Statistical and geometric reconstruction of hominid
crania: Reconstructing australopithecine ontogeny. Ph.D.dissertation,
University of Vienna, Vienna.
Gunz, P., & Harvati, K. (2007). The Neanderthal “chignon”: Variation,
integration and homology. Journal of Human Evolution, 52,
262–274.
Gunz, P., Mitteroecker, P., & Bookstein, F. L. (2005). Semilandmarks in
three dimensions. In D. E. Slice (Ed.), Modern morphometrics in
physical anthropology (pp. 73–98). New York: Kluwer/Plenum.
Gunz, P., Mitteroecker, P., Neubauer, S. Weber, G. W., Bookstein, F. L.
(2009). Principles for the virtual reconstruction of hominin crania.
Journal of Human Evolution, 57(1), 48– 62.
Harvati, K. (2001). The Neanderthal problem: 3-D geometric morphometric models of cranial shape variation within and among species.
Ph.D.dissertation, City University of New York, New York.
Harvati, K., Reddy, D. P., & Marcus, L. F. (2002). Analysis of the posterior cranial profile morphology in Neanderthals and modern
humans using geometric morphometrics. American Journal of
Physical Anthropology, S34, 83.
Hublin, J. J. (1978). Apomorphic characters of Neanderthalian skull and
their phylogenetic interpretation. Comptes Rendus Hebdomadaires
des Séances de L’Académie des Sciences SérieD, 287(10),
923–926.
Hublin, J. J. (1988). Caractères dérivés de la région occipito-­
mastoïdienne chez les Néandertaliens. L’Anatomie, 3, 67–73.
Jelinek, J. (1969). Neanderthal man and Homo sapiens in Central and
Eastern Europe. Current Anthropology, 10(5), 475.
P. Gunz and K. Harvati
Lieberman, D. E. (1995). Testing hypotheses about recent human-­
evolution from skulls – integrating morphology, function, development, and phylogeny. Current Anthropology, 36(2), 159–197.
Lieberman, D. E., Pearson, O. M., & Mowbray, K. M. (2000).
Basicranial influence of overall cranial shape. Journal of Human
Evolution, 38, 291–315.
Mardia, K. V., & Bookstein, F. L. (2000). Statistical assessment of bilateral symmetry of shapes. Biometrika, 87, 285–300.
Mitteroecker, P. (2007). Modularity and Integration. Ph.D. dissertation,
University of Vienna, Vienna.
Mitteroecker, P., & Bookstein, F. L. (2007). The conceptual and statistical relationship between modularity and morphological integration.
Systematic Biology, 56(5), 818–836.
Reddy, D. P., Harvati, K., & Kim, J. (2005). Alternative approaches to
ridge-curve analysis using the example of the Neanderthal occipital
bun. In D. Slice (Ed.), Modern morphometrics in physical anthropology (pp. 99–115). New York: Kluwer/Plenum.
Rohlf, F. J., & Slice, D. (1990). Extensions of the Procrustes method for
the optimal superimposition of landmarks. Systematic Zoology, 39,
40–59.
Smith, F. H. (1982). Upper Pleistocene Hominid evolution in South–
Central Europe: A review of the evidence and analysis of trends.
Current Anthropology, 23, 667–703.
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. 211–250). New York: Liss.
Smith, F. H., Jankovič, I., & Karavanič, I. (2005). The assimilation
model, modern human origins in Europe, and the extinction of
Neandertals. Quaternary International, 137, 7–19.
Trinkaus, E., & LeMay, M. (1982). Occipital bunning among Later
Pleistocene hominids. American Journal of Physical Anthropology,
57, 27–35.
Vlcek, E. (1970). Relations morphologiques des types humains fossiles
de Brno et Cro-Magnon auPleistocene Supérieur d’Europe. In G.
Camps & G. Olivier (Eds.), L’Homme de Cro-Magnon (pp. 59–72).
Paris: Arts et Métiers Graphiques.
Wolpoff, M. H., Hawks, J., Frayer, D., & Hunley, K. (2001). Modern
human ancestry at the peripheries: A test of the replacement theory.
Science, 291, 293–297.
Wrigth, S. (1932). General, group and special size factors. Genetics, 15,
603–619.