AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 136:108–113 (2008)
Brief Communication: Human Cranial Variation Fits
Iterative Founder Effect Model With African Origin
Noreen von Cramon-Taubadel1 and Stephen J. Lycett2*
1
Leverhulme Centre for Human Evolutionary Studies, University of Cambridge, The Henry Wellcome Building,
Fitzwilliam Street, Cambridge CB2 1QH, UK
2
British Academy Centenary Research Project, SACE, University of Liverpool, Hartley Building,
Brownlow Street, Liverpool L69 3BX, UK
KEY WORDS
human origins
cranial diversity; within-group variance; iterative founder effect model; modern
ABSTRACT
Recent studies comparing craniometric
and neutral genetic affinity matrices have concluded
that, on average, human cranial variation fits a model of
neutral expectation. While human craniometric and
genetic data fit a model of isolation by geographic distance, it is not yet clear whether this is due to geographically mediated gene flow or human dispersal events.
Recently, human genetic data have been shown to fit an
iterative founder effect model of dispersal with an African
origin, in line with the out-of-Africa replacement model
for modern human origins, and Manica et al. (Nature 448
(2007) 346–349) have demonstrated that human craniometric data also fit this model. However, in contrast with
the neutral model of cranial evolution suggested by previous studies, Manica et al. (2007) made the a priori
assumption that cranial form has been subject to climati-
cally driven natural selection and therefore correct for climate prior to conducting their analyses. Here we employ
a modified theoretical and methodological approach to
test whether human cranial variability fits the iterative
founder effect model. In contrast with Manica et al.
(2007) we employ size-adjusted craniometric variables,
since climatic factors such as temperature have been
shown to correlate with aspects of cranial size. Despite
these differences, we obtain similar results to those of
Manica et al. (2007), with up to 26% of global within-population craniometric variation being explained by geographic distance from sub-Saharan Africa. Comparative
analyses using non-African origins do not yield significant
results. The implications of these results are discussed in
the light of the modern human origins debate. Am J Phys
Anthropol 136:108–113, 2008. V 2007 Wiley-Liss, Inc.
Recently, a number of studies have assessed the extent
to which modern human cranial diversity patterns fit an
evolutionary model of neutral expectation (e.g. Relethford, 2002; Gonzalez-Jose et al., 2004; Roseman, 2004;
Harvati and Weaver, 2006a,b; Smith et al., 2007). If neutral forces are largely responsible for shaping human
phenotypic diversity patterns, then population affinity
patterns can be employed to infer past population history. Roseman (2004) tested the neutral hypothesis by
comparing human morphological affinity patterns with
microsatellite data and found that, on average, modern
human cranial morphology varied among populations
according to neutral expectation, although some aspects
of cranial shape in high latitude populations may have
been shaped by natural selection. Subsequent analyses
employing 3D geometric morphometric data largely concur with these conclusions. Harvati and Weaver
(2006a,b) show that neurocranial and temporal bone
shape track neutral genetic distances, while facial shape
reflects climate. Smith et al. (2007) confirm the results of
Harvati and Weaver (2006a,b) in finding that temporal
bone morphology reflects neutral rather than selective
patterns of evolution.
It has been demonstrated previously that global
human craniometric and genetic affinity patterns fit a
model of isolation by distance (IBD) (Cavalli-Sforza
et al., 1994; Eller, 1999; Relethford, 2004a,b), which predicts that as the geographic distance increases between
populations, so too will their genetic and phenotypic dissimilarity (Wright, 1943). This is because in cases where
selection has not had a dominant effect on structuring
variance between populations, geographic proximity
mediates potential for migration and thus potential for
gene flow. Relethford (2004b) points out that a fit to an
IBD model does not make clear the underlying evolutionary process, and does not make possible the distinction
between a model of long-distance gene flow mediated by
geographic distances or one of a global dispersal event.
Manica et al. (2007) demonstrated recently that modern human cranial diversity patterns fit a dispersal
model of iterative founder effects (repeated bottlenecking) with an African origin. Their results are in accordance with similar results obtained using neutral autosomal microsatellite markers, which show that population
heterozygosity fits a stepping-stone model of dispersal from an African origin (Prugnolle et al., 2005;
C 2007
V
WILEY-LISS, INC.
C
Grant sponsors: St. John’s College, University of Cambridge;
Gates Trust Scholarship; British Academy Centenary Research Project: ‘‘Lucy to Language’’.
*Correspondence to: Stephen J. Lycett, School of Archaeology,
Classics and Egyptology, Hartley Building, Brownlow Street, University of Liverpool, Liverpool L69 3BX, UK.
E-mail: [email protected]
Received 9 August 2007; accepted 7 November 2007
DOI 10.1002/ajpa.20775
Published online 27 December 2007 in Wiley InterScience
(www.interscience.wiley.com).
109
HUMAN CRANIAL VARIATION
TABLE 1. Twenty-eight human populations sampled,
including samples sizes, within-group variances, and
geographic co-ordinates
Region
Africa
Europe
Asia
Oceania
Americas
Polynesia
Population
Sample size
WGV
Latitude,
longitude
Teita
Egypt
Zulu
San
Dogon
Zalavar
Berg
Norse
Andamanese
Buriats
Hainan
Anyang
Philippines
Atayal
South Japan
North Japan
Ainu
Guam
Australian
Tolai
Tasmania
Santa Cruz
Arikara
Peruvian
Inugsuk
Mokapu
Moriori
Easter Island
33
58
55
41
47
53
56
55
35
55
45
42
50
29
50
55
48
30
52
56
45
51
42
55
53
51
57
49
0.85
0.64
0.77
1.01
0.83
0.64
0.74
0.66
0.81
0.76
0.68
0.75
0.70
0.70
0.73
0.80
0.63
0.58
0.56
0.60
0.70
0.59
0.59
0.62
0.77
0.61
0.65
0.60
3.3S, 38.6E
30.0N, 31.5E
28.5S, 30.2E
22.4S, 19.0E
14.5N, 0.5E
46.8N, 17.1E
46.8N, 13.2E
59.9N, 10.8E
12.4N, 92.8E
52.8N, 108.7E
19.4N, 110.2E
36.0N, 114.3E
14.6N, 121.0E
24.3N, 121.1E
33.0N, 131.0E
43.0N, 141.0E
43.5N, 142.0E
13.4N, 144.8E
35.0S, 139.8E
5.4S, 151.0E
41.7S, 146.1E
34.0N, 119.8W
44.7N, 100.1W
12.5S, 75.9W
61.7N, 48.7W
21.5N, 157.9W
43.9S, 176.5W
27.0S, 109.3W
Ramachandran et al., 2005; Liu et al., 2006). Moreover,
Manica et al. (2007) test a model of multiregional origins
for modern humans but fail to find evidence of a second
non-African origin capable of explaining any residual
variation in the data.
Here we employ Howells’ (1996) craniometric dataset
to further investigate whether global human cranial diversity patterns can be explained by a dispersal event
out of Africa, following an iterative founder effect model.
This study differs from that of Manica et al. (2007) in
both theoretical expectation and methodological
approach. In direct contrast with the neutral model for
human cranial form purported on the basis of previous
analyses (e.g., Roseman, 2004; Harvati and Weaver,
2006b), Manica et al. (2007) make an a priori assumption that cranial dimensions are correlated with climate,
implying that natural selection has played a significant
role in the evolution of the modern human cranium.
Accordingly, Manica et al. (2007) correct for within-population climatic variability prior to conducting their analyses. Here we do not assume that climate has had a significant impact on overall cranial shape, but acknowledge that climatic variables such as mean annual
temperature have been shown to be correlated with
aspects of cranial size (Harvati and Weaver, 2006a;
Smith et al., 2007) indicating a conformation (at least in
part) of the cranium to Bergmann’s rule. As in Manica
et al. (2007), a significant inverse relationship between
geographic distance from sub-Saharan Africa (i.e. along
a hypothesized dispersal route) and within-population
variability is taken as support for the iterative founder
effect model of dispersal. For comparative purposes,
three non-African origins were also tested.
TABLE 2. Geographic co-ordinates for the three African and
three non-African dispersal origin points and for
the five way-points used to calculate the geographic
distances along dispersal routes
Start points
Waypoints
Addis Ababa
Central Africa
Southern Africa
Tel Aviv
Delhi
Beijing
Cairo
Istanbul
Phnom Penh
Anadyr
Panama
9.0N, 38.0E
0.0N, 25.0E
20.0S, 25.0E
32.0N, 34.8E
28.5N, 77.0E
40.0N, 116.4E
30.0N, 31.0E
41.0N, 28.0E
11.0N, 104.0E
64.0N, 177.0E
13.5N, 86.2W
MATERIALS AND METHODS
Cranial measurement data were obtained from W.W.
Howells’ extensive and freely-available database (Howells,
1973, 1996). These data comprise 57 cranial variables for
28 globally distributed human populations, including
three Polynesian populations. Manica et al. (2007)
removed oceanic populations from their study due to the
uncertain settlement history of these islands. Here, to
produce results comparable with those of Manica et al.
(2007), all analyses were first conducted with the Polynesian populations excluded and subsequently all analyses
were repeated with Polynesians included. All of the populations sampled by Howells contain male specimens but
not all contain data for females. Therefore, to avoid the
potentially confounding effects of sexual dimorphism,
only male data were used. Table 1 provides population
names, sample sizes, within-population variance, and geographic co-ordinates for all craniometric data employed.
To remove the potentially confounding influence of climate on cranial size (Harvati and Weaver, 2006a; Smith
et al., 2007) all raw cranial measurements were sizeadjusted by dividing each measurement by the geometric
mean of all measurements for that individual (Jungers
et al., 1995). This method removes isometric scaling,
thereby maintaining the overall shape of the object
(Falsetti et al., 1993). The average within-group variance
across all 57 cranial characters for each of the 28 populations was calculated in RMET 5.0 (Relethford and
Blangero, 1990).
Longitude and latitude co-ordinates were estimated for
all populations (see Table 1), following the map and population descriptions provided by Howells (1989). Three
dispersal origins were chosen in sub-Saharan Africa (Table 2). The first of these, Addis Ababa (Ethiopia), was
chosen following Prugnolle et al. (2005), as the location
where some of the oldest anatomically modern human
remains have been discovered (White et al., 2003). Ramachandran et al. (2005) sought the most likely origin of a
human expansion from a lattice of over 4,000 possible origin points and were able to identify an area in central
sub-Saharan Africa. This empirical observation was confirmed by Manica et al. (2007) using a similar analysis
of the same genetic data. They also identified a subSaharan region giving the strongest relationship
between within-population phenotypic variance and geographic distance, which differed from the genetic estimate by including South Africa amongst the likely origins. Therefore two additional African origin points were
chosen for this study—one from central equatorial
American Journal of Physical Anthropology
110
N.
VON
CRAMON-TAUBADEL AND S.J. LYCETT
Fig. 1. Regression of average within-group variance of 57 size-adjusted craniometric variables on great circle distance (km)
from African dispersal origins. A: Addis Ababa (Ethiopia). B: Central Africa (Democratic Republic of the Congo). C: Southern Africa
(Botswana). Diamonds 5 African populations. Open squares 5 European populations. Triangles 5 Asian populations. Stars 5 Oceanic populations. Open circles 5 American populations.
African (within the Democratic Republic of the Congo)
and another from southern Africa (Botswana) (see Table 2
for co-ordinates). For comparative purposes, three nonAfrican start points were also tested. These comprised Tel
Aviv (Israel), Delhi (India) and Beijing (China) (Table 2).
Following Ramachandran et al. (2005), hypothetical
dispersal routes were estimated using way-points
between continents (Table 2). These were comprised of:
1) Cairo, Egypt (entry/exit point to Africa); 2) Istanbul,
Turkey (entry/exit point to Europe); 3) Phnom Penh,
Cambodia (entry/exit point to Oceania including Polynesia); 4) Anadyr, Russia (entry/exit point to New World);
5) Panama (entry/exit point to South America). Geographic distances between origin points and all populations were calculated as great circle distances based on
the haversine (Sinnott, 1984). Great circle distances take
account of the earth’s curvature and are appropriate
measures of distance when covering large geographic
areas. Least-squares regressions were performed in
SPSS 12.0.1 using the three African and three non-African dispersal origins. Analyses were repeated following
removal of the Polynesians. In each case the independent variable of average within-group variance was
regressed on the dependent variable of geographic distance from the point of dispersal.
RESULTS
Figure 1 shows the regression results for the three
African dispersal origins when the Polynesians were
American Journal of Physical Anthropology
excluded. All origin points yield a significant inverse
linear relationship between within-population phenotypic variance and geographic distance, as predicted by
the iterative founder effect model. Between 19 and 26%
of the within-group variance was explained by distance
from an African origin, with the southern African origin yielding the best fit to the model. In contrast, Figure 2 shows the regression results taking a non-African
dispersal origin, none of which yield a significant
result. Although not all possible alternative origins
have been tested, these results concur with those of
Manica et al. (2007), which found no origin outside
Africa capable of explaining any residual variation in
the data. Table 3 provides the comparable results when
all 28 populations, including the Polynesians, were
employed. Inclusion of the Polynesian populations
strengthens the fit of the data to an iterative founder
effect model with between 24 and 31% of within-group
variance being explained by distance from an African
origin. Therefore exclusion of the Polynesian populations can be considered a conservative analytical approach to this analysis.
Manica et al. (2007) make the assumption that climatically driven natural selection has shaped some
aspects of cranial variability, and thus take a conservative approach to their data by correcting for climatic
variation prior to analysis. Relethford (2004a) has
shown that climatic variation only accounts for a small
amount of the residual craniometric variation when
compared with geographic distance, indicating that cli-
111
HUMAN CRANIAL VARIATION
Fig. 2. Regression of average within-group variance of 57 size-adjusted craniometric variables on great circle distance (km)
from non-African dispersal origins. A: Tel Aviv (Israel). B: Delhi (India). C: Beijing (China). Diamonds 5 African populations.
Open squares 5 European populations. Triangles 5 Asian populations. Stars 5 Oceanic populations. Open circles 5 American
populations.
TABLE 3. Regression of average within-group variance of 57
size-adjusted craniometric variables on great circle distance
(km) from three African dispersal origins and three
non-African dispersal origins, when the three Polynesian
populations were included in the analysis
Dispersal origins
African
Non-African
R-values (P-values)
Addis Ababa
Central Africa
Southern Africa
Tel Aviv
Delhi
Beijing
20.49
20.52
20.56
20.41
20.23
0.01
(0.008)
(0.004)
(0.002)
(0.030)
(0.230)
(0.950)
mate does not obscure the overall pattern of IBD on a
global scale. Although some studies (Relethford, 2004a;
Roseman, 2004; Roseman and Weaver, 2004; Harvati
and Weaver, 2006b) have shown that certain aspects of
morphology do correlate with climatic variation, it has
been demonstrated that this effect is strongest in high
latitude populations, who are subject to the greatest
extremes of climate (Roseman, 2004; Harvati and
Weaver, 2006b). Indeed, Harvati and Weaver (2006b)
found that when populations from high latitudes were
removed from the analysis, no cranial features were
found to correlate with climate. To evaluate the impact
of the high latitude populations on our analysis, we
removed the Inugsuk (Greenland Inuit) from the dataset
and reran the analysis with the southern African startpoint, which had yielded the strongest R2-value previously (when Polynesians were excluded). When the Inuit
population was removed, the strength of the relationship between geographic distance from Africa and
within-population variance increased significantly (R2 5
0.42, P 5 0.001).
DISCUSSION
The use of an additional dataset of global craniometric
variation corroborates the results of Manica et al. (2007)
in demonstrating that modern human craniometric variance patterns fit a model of iterative founder effects
along a dispersal route from an African origin. Despite
the theoretical and methodological differences, our
results are remarkably similar to those of Manica et al.
(2007) who found that distance from Africa could account
on average for 19–25% of the variation in craniometric
traits. We find an African origin dispersal model could
explain at least 19–26% of within-population variation,
when the Polynesian populations were excluded.
Although we do not explicitly correct for climatic variation, by removing isometric scaling prior to analysis, the
effects of environmentally driven selection on size (Beals
et al., 1983; Roseman, 2004; Harvati and Weaver, 2006a;
Smith et al., 2007) were mitigated. However, removing
high latitude populations had a more dramatic effect on
the results, greatly increasing the explanatory power of
the model. This presumably reflects the environmentally
induced adaptation to extremes of temperature experienced by such populations.
A neutral model for the evolution of cranial variation
in humans would predict that phenotypic affinity patterns should match closely those of neutral genetics.
American Journal of Physical Anthropology
112
N.
VON
CRAMON-TAUBADEL AND S.J. LYCETT
Thus the fit of neutral genetic data with an African dispersal model (Prugnolle et al., 2005; Ramachandran
et al., 2005; Liu et al., 2006) indicates that craniometric
variation should also fit a dispersal model. In the case
of the genetic data, however, over 87% of population heterozygosity can be explained by the model, contrasted
with 25% of within-population variance for craniometric traits. However, as pointed out by Manica et al.
(2007), cranial variation is only partly determined by
genetic variation, with average heritability values for
craniometric traits commonly cited as h2 5 0.55 (Relethford, 1994; Relethford and Harpending, 1994). Therefore, once heritability and potential diversifying selection are accounted for, these lower estimates for phenotypic data are to be expected. In addition, the results of
Harvati and Weaver (2006a,b) indicate that individual
regions of the human cranium may be reflecting neutral
evolutionary history more effectively than others. Therefore, future studies may benefit from assessing whether
individual cranial regions provide a better fit with the
iterative founder effect model than the cranium as a
whole.
Some have claimed that with multiple lines of genetic,
phenotypic, and archeological evidence pointing toward
an African origin for modern humans (e.g. Lahr, 1994;
Stringer, 2002), one end of the polarized ‘‘multiregional’’
versus ‘‘out of Africa’’ debate may have been satisfactorily rejected in favor of the latter (Foley and Lahr,
2004). Others, however, contend that rumors of the demise of the multiregional model may have been greatly
exaggerated (e.g. Relethford, 1998; Templeton, 2007).
According to such workers, alternative hypotheses of
modern human origins have not been falsified, resulting
in a situation where genetic data are compatible with a
number of working hypotheses (Templeton, 2007). As
Relethford (2004b) points out, a fit with an IBD model
does not allow for the distinction to be made between a
long-term pattern of geographically mediated gene flow
and a pattern of demographic migration. Therefore, a fit
with an IBD model cannot in itself inform the debate
regarding modern human origins. All that can be concluded with some certainty is that human craniometric
and neutral genetic data is hierarchically structured
according to the geographic distance between populations. The results of this study and those of Manica
et al. (2007) inform the modern human origins debate in
so far as they are compatible with the hypothesis that
Africa is the source of all modern human genetic diversity (Jorde et al., 1998, 2000; Yu et al., 2002), and that
the initial dispersal of modern humans proceeded as a
series of repeated bottlenecking events, as humans
spread around the globe.
ACKNOWLEDGMENTS
We are grateful to Christopher Ruff, the associate
editor and two anonymous reviewers for their helpful
and constructive comments on an earlier draft of this
manuscript.
LITERATURE CITED
Beals KL, Smith CL, Dodd SM. 1983. Climate and the evolution
of brachycephalization. Am J Phys Anthropol 62:425–437.
Cavalli-Sforza LL, Menozzi P, Piazza A. 1994. The history and
geography of human genes. Princeton: Princeton University
Press.
American Journal of Physical Anthropology
Eller E. 1999. Population substructure and isolation by distance
in three continental regions. Am J Phys Anthropol 108:147–
159.
Falsetti AB, Jungers WL, Cole TM III. 1993. Morphometrics of
the Callitrichid forelimb: a case study in size and shape. Int J
Primatol 14:551–572.
Foley RA, Lahr MM. 2004. Modern human origins—why it’s
time to move on. In: Jobling MA, Hurles ME, Tyler-Smith C,
editors. Human evolutionary genetics: origins, peoples and
disease. New York: Garland. p 249–250.
Gonzalez-Jose R, van der Molen S, Gonzalez-Jose E, Hernandez
M. 2004. Patterns of phenotypic covariation and correlation in
modern humans as viewed from morphological integration.
Am J Phys Anthropol 123:69–77.
Harvati K, Weaver TD. 2006a. Reliability of cranial morphology
in reconstructing Neanderthal phylogeny. In: Harvarti K,
Harrison TL, editors. Neanderthals revisited: new approaches
and perspectives. Dordrecht: Springer. p 239–254.
Harvati K, Weaver TD. 2006b. Human cranial anatomy and the
differential preservation of population history and climate signatures. Anat Rec A 288:1225–1233.
Howells WW. 1973. Cranial variation in man: a study by multivariate analysis of patterns of difference among recent human
populations. Cambridge: Harvard University Press.
Howells WW. 1989. Skull shapes and the map. Craniometric
analyses in the dispersion of modern Homo. Cambridge: Harvard University Press.
Howells WW. 1996. Howells’ craniometric data on the internet.
Am J Phys Anthropol 101:441–442.
Jorde LB, Bamshad M, Rogers A. 1998. Using mitochondrial
and nuclear DNA markers to reconstruct human evolution.
Bioessays 20:126–136.
Jorde LB, Watkins WS, Bamshad MJ, Dixon ME, Ricker CE,
Seielstad MT, Batzer MA. 2000. The distribution of human
genetic diversity: a comparison of mitochondrial, autosomal,
and Y-chromosome data. Am J Hum Genet 66:979–988.
Jungers WL, Falsetti AB, Wall CE. 1995. Shape, relative size
and size-adjustments in morphometrics. Yearb Phys Anthropol 38:137–161.
Lahr MM. 1994. The multiregional model of modern human origins: a reassessment of it’s morphological basis. J Hum Evol
26:23–56.
Liu H, Prugnolle F, Manica A, Balloux F. 2006. A geographically
explicit genetic model of worldwide human-settlement history.
Am J Hum Genet 79:230–237.
Manica A, Amos W, Balloux F, Hanihara T. 2007. The effect of
ancient population bottlenecks on human phenotypic variation. Nature 448:346–349.
Prugnolle F, Manica A, Balloux F. 2005. Geography predicts
neutral genetic diversity of human populations. Curr Biol
15:R159–R160.
Ramachandran S, Deshpande O, Roseman CC, Rosenberg NA,
Feldman MW, Cavalli-Sforza LL. 2005. Support from the relationship of genetic and geographic distance in human populations for the serial founder effect originating in Africa. Proc
Natl Acad Sci USA 102:15942–15947.
Relethford JH. 1994. Craniometric variation among modern
human populations. Am J Phys Anthropol 95:53–62.
Relethford JH. 1998. Genetics of modern human origins and diversity. Ann Rev Anthropol 27:1–23.
Relethford JH. 2002. Apportionment of global human genetic diversity based on craniometrics and skin color. Am J Phys
Anthropol 118:393–398.
Relethford JH. 2004a. Boas and beyond: migration and craniometric variation. Am J Hum Biol 16:379–386.
Relethford JH. 2004b. Global patterns of isolation by distance
based on genetic and morphological data. Hum Biol 76:499–
513.
Relethford JH, Blangero J. 1990. Detection of differential gene
flow from patterns of quantitative variation. Hum Biol 62:5–
25.
Relethford JH, Harpending HC. 1994. Craniometric variation,
genetic theory, and modern human origins. Am J Phys
Anthropol 95:249–270.
HUMAN CRANIAL VARIATION
Roseman CC. 2004. Detecting interregionally diversifying natural selection on modern human cranial form by using matched
molecular and morphometric data. Proc Natl Acad Sci 101:
12824–12829.
Roseman CC, Weaver TD. 2004. Multivariate apportionment of
global human craniometric diversity. Am J Phys Anthropol
125:257–263.
Sinnott RW. 1984. Virtues of the haversine. Sky Telescope 68:
159.
Smith HF, Terhune CE, Lockwood CA. 2007. Genetic, geographic, and environmental correlates of human temporal
bone variation. Am J Phys Anthropol 134:312–322.
113
Stringer C. 2002. Modern human origins: progress and prospects. Philos Trans R Soc Lond B 357:563–579.
Templeton AR. 2007. Genetics and recent human evolution.
Evolution 61:1507–1519.
White TD, Asfaw B, deGusta D, Gilbert H, Richards GD, Suwa
G, Howell FC. 2003. Pleistocene Homo sapiens from Middle
Awash, Ethiopia. Nature 423:742–747.
Wright S. 1943. Isolation by distance. Genetics 28:114–138.
Yu N, Chen F-C, Ota S, Jorde LB, Pamilo P, Patthy L, Ramsay
M, Jenkins T, Shyue S-K, Li W-H. 2002. Larger genetic differences within Africans than between Africans and Eurasians.
Genetics 161:269–274.
American Journal of Physical Anthropology