Old World sources of the first New World human
inhabitants: A comparative craniofacial view
C. Loring Brace*
†
, A. Russell Nelson*
‡
, Noriko Seguchi*, Hiroaki Oe
§
, Leslie Sering*, Pan Qifeng
¶
, Li Yongyi
i
,
and Dashtseveg Tumen**
*Museum of Anthropology, University of Michigan, 1109 Geddes Avenue, Ann Arbor, MI 48109;
‡
Department of Anthropology, University of
Wyoming, Laramie, WY 82071;
§
Department of Statistics, University of Michigan, Ann Arbor, MI 48109;
¶
Institute of Archaeology, Chinese
Academy of Social Sciences, 27 Wangfujing Dajie, Beijing 100710, China;
i
Department of Anatomy, Chengdu College of Traditional
Chinese Medicine, 13 Xing Lo Road, Chengdu, Sichuan, People’s Republic of China; and **Department of Anthropology,
Mongolian Academy of Sciences, Ulaanbaatar-51, Mongolia
Communicated by Kent V. Flannery, University of Michigan, Ann Arbor, MI, June 18, 2001 (received for review January 2, 2001)
Human craniofacial data were used to assess the similarities and
differences between recent and prehistoric Old World samples,
and between these samples and a similar representation of
samples from the New World. The data were analyzed by the
neighbor-joining clustering procedure, assisted by bootstrap-
ping and by canonical discriminant analysis score plots. The first
entrants to the Western Hemisphere of maybe 15,000 years ago
gave rise to the continuing native inhabitants south of the
U.S.–Canadian border. These show no close association with any
knownmainland Asian population. Instead they show ties to the
Ainu of Hokkaido and their Jomon predecessors in prehistoric
Japan and to the Polynesians of remoteOceania. All of these also
have ties to the Pleistocene and recent inhabitants of Europe and
may represent an extension from a Late Pleistocene continuum
of people across the northern fringe of theOldWorld.With roots
in both the northwest and the northeast, these people can be
described as Eurasian. The route of entry to the NewWorld was
at the northwestern edge. In contrast, the Inuit (Eskimo), the
Aleut, and the Na-Dene speakers who had penetrated as far as
the American Southwest within the last 1,000 years show more
similarities to the mainland populations of East Asia. Although
both the earlier and later arrivals in the New World show a
mixture of traits characteristic of the northern edge ofOldWorld
occupation and the Chinese core of mainland Asia, the propor-
tion of the latter is greater for the more recent entrants.
New World origins u craniometrics u prehistoric population relationships
T
he first OldWorld travelers to record their observations on
the people they found living in the Western Hemisphere
took it for granted that all humans ultimately descended from
a single original pair as described in the Judaeo-Christian
Bible. The exact location of that supposed initial Eden was not
known, but it was generally assumed to have been somewhere
in the Middle East of the Old World. Subsequently, thoughtful
observers, such as Fray Jose´ de Acosta in the late sixteenth
century and Thomas Jefferson in the eighteenth, realized that
eastern Asia had to be considered the most probable imme-
diate source when questions concerning the locus of New
World human origins were raised (1). Some assumed that there
was a link with wandering ‘‘Tartars’’ or the legendary ‘‘Ten
Lost Tribes of Israel’’ (2), but, as faith waned in the Bible as
a source of scientific information, the suggestion was proposed
that human beings as well as the rest of animate nature were
separately created in each of the geographic provinces of the
world (3, 4). Charles Darwin then developed a non-Biblical
explanation for the common origin for all living humans and
made the case that the ultimate locus of human ancestry most
probably was to be sought in Africa (5).
The fact that there is no skin color cline from the Arctic to the
Equator in the native inhabitants of the New World indicates
that occupation of the western hemisphere is of too shallow a
time depth to have produced such an adaptive gradient. Skin
color, however, is so different from that found in sub-Saharan
Africa that it clearly indicates long residence in the temperate
rather than the tropical latitudes of the Old World. If the
gradient in Australia is a product of some 60,000 years of
occupation, then the picture north of the Equator had to have
taken approximately three times as long in situ for selection to
have produced the picture visible in the temperate zone (6).
Genetic (7) and archaeological evidence (8) supports a northeast
Asian source for the first human inhabitants of the Western
Hemisphere. Some interpretations have suggested that the dis-
tribution of linguistically identifiable groups in the New World
may have been the result of separate prehistoric population
movements into the Western Hemisphere (9, 10).
From the sparse archaeological evidence and complementary
molecular genetic data from living populations, an initial date of
ca. 15,000 years can be postulated (11). Questions concerning the
initial human settlement of the New World have involved such
matters as the initial date of entry, route of access, whether there
was a single or several dissimilar waves of people, and how these
are genetically related both to living American Indian groups as
well as to Asian and Pacific populations and possibly to other Old
World peoples (12–14). Issues of geology, archaeology, and legal
ownership are all involved (15–19).
Materials and Methods
Morphometric Comparisons. Recently these questions have been
investigated by comparing patterns of the geographical distri-
bution of human genetic and morphological features (20–26).
Metric variables record inherited differences in cranial and
facial form by documenting minor variations in the configu-
ration of suture placement, length, and other details in the
construction of the cranial vault and face. The various con-
figurations of craniofacial form cluster regionally and are not
distributed in clinal fashion related to the intensity of different
selective force gradients. Furthermore, once established, con-
figurations of facial form appear to remain relatively stable
over considerable spans of time (27–29). Consequently, al-
though details of facial form are distributed as regional
groupings that may be viewed in the context of implied levels
of biological relationship, these considerations must be made
at a level well below that of the species. The human genotypic
web is a subspecific, and therefore genetically open, system.
None of the earlier human groups sampled in our data set, if
they were all alive at the same time, would have been repro-
ductively isolated in a fashion analogous to that of different
species in a cladistic scheme. Because of this fact, the metri-
cally indicated relationships depicted here, although of heu-
†
To whom reprint requests should be addressed. E-mail: clbrace@umich.edu.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
www.pnas.orgycgiydoiy10.1073ypnas.171305898 PNAS u August 14, 2001 u vol. 98 u no. 17 u 10017–10022
A
N
T
H
R
O
P
O
L
O
G
Y

ristic value in the examination of similarities and differences
between groups of people across regions and through time, do
not constitute a taxonomy in the strict sense of the term (30).
Here we integrate the results of the analysis of craniometric
data showing the similarities and differences between prehistoric
and recent samples representing the major regions of the Old
World, the Pacific, and the New World. Our analysis is based on
a set of 21 craniofacial measurements made of collections of
human crania in the samples tested. These measurements are
listed in Table 1. Preliminary studies (31–33) used C-scores and
Euclidean distance clustering procedures to compare New
World samples with a selected number of OldWorld groups. For
the present analysis, we increased the number of Old World
samples and added prehistoric material for comparison. The
groups used, along with their sample sizes, are recorded in
Table 2.
After each bootstrap resampling of the original data (34–37),
Mahalanobis distances were calculated for all pairs of the groups
being compared. Then the neighbor-joining method of dendro-
gram building was used to display their similarities and differ-
ences (38). To quantify the relationships of the groups being
compared, stepwise canonical discriminant function scores were
calculated (39). To use data from the prehistoric samples where
individuals with complete sets of measurements were hard to
find, a multivariate procedure for imputing missing values was
used (40).
Results
The resemblances and distinctions of samples representing the
main geographic regions of the Old World are depicted in the
neighbor-joining dendrogram displayed in Fig. 1. Also repre-
sented is the placement of the samples in relation to each other
as determined by the first two discriminant functions. A com-
posite of European Upper Paleolithic specimens from Czecho-
slovakia to France between 20 and 30 kya and a sampling of four
European Neanderthals (from Gibraltar, France, and Italy) of
'50 kya are also included.
Prehistoric and Recent Components of the Old World. The Nean-
derthal twig of the dendrogram (Fig. 1A) is separated from the
others by a greater distance than that discernible between any
other possible pair. On the other hand, no matter what samples
are used, it always links to the European twig before any other.
This link is confirmed by the plot based on the first two
canonical discriminant function scores (Fig. 1B). Only when a
third discriminant function (accounting for 9.6% of total
variance) is used (or, alternatively, a third principal compo-
nent) do Neanderthals separate noticeably from recent Euro-
peans. The separation of the northeast Africans as well as the
European and South Asian samples from sub-Saharan Africa
is compatible with the picture of genetic attenuation when the
core of the African continent is compared with its northeast-
ern edge and to extra-African peoples (41). The tie between
Africa and Australo-Melanesia also has parallels at the mo-
lecular genetic level, where anAfrican degree of haplotype and
microsatellite diversity has echoes in the southern part of
South Asia and over into Melanesia and Australia (42–45).
One explanation is that this haplotype dates back to the initial
human spread fromAfrica across the tropics of the OldWorld.
The morphological and genetic distinctions of populations at
the northwestern and northeastern edges of the Old World
would then be the result of the subsequent northward spread
of peripheral extensions from the continuous tropical corpus
of the Old World’s main Paleolithic population concentration.
Neither the genetic nor the morphological picture can give us
any definitive estimate of the timing of that initial occupation
of the tropics or the subsequent northward extensions.
The archaeological record, however, does provide some sug-
gestions concerning population movements to the north. Such
movements are crucial for establishing a temporal framework for
the occupation of the northernmost edge of habitation in theOld
World and for the extension of that occupation into the Western
Hemisphere. The control of fire made occupation north of the
tropics possible before the penultimate glaciation (46). At the
western third of the temperate zone, a subsequent extension
farther north was accompanied by the use of hafted projectile
(Levallois) points—ultimately derived from the African Middle
Stone Age with roots in the Acheulean more than 200,000 years
ago (47). The subsequent distribution of the use of such tools
eastward across the northern edge of human habitation to
Mongolia and Siberia, but not to China and Japan or down
toward Southeast Asia (8, 48), suggests an actual continuity
between the makers, whether as the result of gene flow or an
actual movement of people from the west.
Table 1. Craniofacial measures in the UMMA dataset (27)
Var. no. Description
1 Nasal height
2 Nasal bone height
3 Piriform aperture height
4 Nasion prosthion length
5 Nasion basion
6 Basion prosthion
7 Superior nasal bone width
8 Simotic width
9 Inferior nasal bone width
10 Nasal breadth
11 Simotic subtense
12 Inferior simotic subtense
13 FOW (fronto-orbital width) subtense at nasion
14 MOW (mid-orbital width) subtense at rhinion
15 Bizygomatic breadth
16 Glabella opisthocranion
17 Maximum cranial breadth
18 Basion bregma
19 Basion rhinion
20 Width at 13 (fmt fmt)
21 Width at 14
Table 2. Samples and Ns used in the present analysis
Population Female N Male N Total
Africa 52 54 106
Ainu 23 33 56
Athabaskan 27 21 166
AustraliayMelanesia 74 92 166
China 139 166 305
Europe 152 190 342
Europe Upper Paleo 4 3 7
InuityAleut 76 84 160
Jomon 6 3 9
Mongol 21 29 50
Mongolia Bronze 25 29 54
Neanderthal 2 2 4
Polynesia 75 64 139
PuebloyMexicoyPeru 44 46 90
SomaliyNubia 15 38 53
South Asia 46 68 114
Southeast Asia 47 70 117
U.S.–Canada border 24 23 47
Upper CaveyMinato 0 3 3
10018 u www.pnas.orgycgiydoiy10.1073ypnas.171305898 Brace et al.