Variation Among Early North American Crania
R.L. Jantz1
*
and Douglas W. Owsley2
1
Department of Anthropology, University of Tennessee, Knoxville, Tennessee 37996-0720
2
Department of Anthropology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560
KEYWORDS Paleoamericans; cranial morphometrics; New World
ABSTRACT The limited morphometric work on early
American crania to date has treated them asasingle,
temporally defined group. This paper addresses the ques-
tionofwhether there is significant variability among an-
cientAmerican crania. A sample of 11 crania (Spirit Cave,
Wizards Beach, Browns Valley, Pelican Rapids, Prospect,
Wet Gravelmale, Wet Gravel female, Medicine Crow,
Turin, Lime Creek, and Swanson Lake) dating from the
early to mid Holocene was available. Some have recent
accelerator mass spectrometry (AMS) dates, while others
are dated geologically or archaeologically. All are in excess
of 4500BP, and most are 7000 BP or older. Measurements
follow the definitions of Howells[(1973) Cranial variation
in man, Cambridge: Harvard University). Some crania are
incomplete, but 22 measurements were common to all
fossils. Cranial variation was examined by calculating the
Mahalanobis distance between each pair of fossils, usinga
pooled within sample covariance matrix estimated from
the data of Howells. The distance relationships among
crania suggest the presence of at least three distinct
groups: 1) a middle Archaic Plains group (Turinand Med-
icineCrow), 2) a Paleo/Early Archaic Great Lakes/Plains
group (Browns Valley, Pelican Rapids, Lime Creek), and
3) a spatially and temporally heterogeneous group that
includes the Great Basin/Pacific Coast (Spirit Cave, Wiz-
ardsBeach, Prospect) and Nebraska (Wet Gravel specimens
and Swanson Lake).
These crania were also compared to Howells'worldwide
recent sample, which was expanded by including six ad-
ditionalAmerican Indian samples. None of the fossils,
except for the Wet Gravel male, shows any particular
affinity to recent Native Americans; their greatest simi-
laritiesarewith Europe, Polynesia, or East Asia. Several
crania would be atypical in any recent population for
which we have data. Browns Valley, Pelican Rapids, and
Lime Creekarethemost distinctive. They provide evi-
denceforthe presence of an early population that bears no
similarity to the morphometric pattern of recent American
Indians or even to crania of comparable date in other
regions of the continent.
The heterogeneity among early American crania makes
it inadvisable to pool them for purposes of morphometric
analysis. Whether this heterogeneity results from differ-
entearly migrations or one highly differentiated popula-
tioncannotbe established from our data. Our results are
inconsistent with hypotheses of an ancestor-descendent
relationship between early and late Holocene American
populations. They suggest that the pattern of cranial vari-
ationisofrecent origin, at least in the Plains region. Am
JPhysAnthropol 114:146-155,2001.
©
2001 Wiley-Liss, Inc.
Native NorthAmerican populations have vari-
ouslybeenviewed as biologically uniform, or ex-
tremelydiverse. Stewart (1973) emphasized the
phenotypic uniformity, while Hooton (1930) saw
Caucasoid, Melanesoid, and Negritoid"types"in the
American population. Recent work with classical
markers (Cavalli-Sforzaetal., 1994), dentition
(Haydenblit, 1996), and anthropometry (Ousley,
1995) has demonstrated that modern native popula-
tionsarestrongly differentiated. It is obviously im-
portantto understand the extent and patterning of
variation among New World populations, since this
context of diversity is the framework against which
models for the peopling of the New World must be
tested.
A puzzling aspect of the discussion concerning the
origin and history of New World populations is how
little the study of ancient crania has contributed.
This lack of emphasis is in large part due to the
influence of Hrdlicka (1937a,b), who repeatedly attempted
to show that the morphology of supposed
early American crania fell within the range of vari-
ationofrecent Indians, and could not be distin-
guishedfromthem (Owsleyand Jantz, 1999a). Fortunately,
a renewed interest in the crania of early
Americans has occurred, using a modern analytical
framework based on multivariate statistics. Initial
studies have shown that thecraniometric pattern
departs from contemporary American Indians, often
in the direction of European or Southern Pacific
Grant sponsor: Calhoun Foundation; Grant sponsor: National Mu-
seumofNatural History.
We thank Roberta Hall for inviting us to participate in the AAPA
symposium"Biological Variation and Population Origins in the Amer-
icasandAustralia,"at which aversion of this paper was presented in
1998.
*Correspondence to: R.L. Jantz, Department of Anthropology, Uni-
versityof Tennessee, Knoxville, TN 37996-0720.
E-mail: jantz@utkux.utk. edu
Received 7December 1998; accepted 21 September 2000.
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 114:146-155 (2001)
©
2001 WILEY-LISS, INC.
groups (Steeleand Powell, 1992,1994), or even Af-
ricansand Australians (Neves and Pucciarelli, 1991;
Neves et al., 1996).
What has not yet been approached in asystematic
manner is variation among Paleoamerican fossils.
This issue is difficult to address, because the sam-
plesare individual specimens, and standard statistical
approaches designed to assess variation among
samples are inappropriate. This paper addresses
that issue as a preliminary step toward understand-
ingwhetherthe early New World populations were
differentiated. If early populations are relatively homogeneous,
it places a time constraint on the vari-
ationobservedin present populations and forces us
to consider later events for causation. On the other
hand, if the early populations were strongly differentiated,
then the question becomes one of how the
ancient diversity relates to recent diversity, and
whether continuity with recent populations can be
established.
MATERIALS AND METHODS
Crania
Ancient crania available for this analysis are
shown in Table 1. The sample includes those early
crania that have been measured using the measure-
mentprotocolofHowells (1973) (see Measurements,
below). The date range includes specimens that are
considered by most archaeologists to be marginally
"Paleoamerican,"or definitely within the Early Ar-
chaicperiod. We did not require that chronometric
dates be firmly established, only that the skull falls
into the early Holocene, not later than 4500 BP.
Several specimens have recent accelerator mass
spectrometry (AMS) dates and are accurately dated
(Turin, Spirit Cave, Wizards Beach, Browns Valley,
and Pelican Rapids). Others have attributed geolog-
icaldates, such as the Wet Gravel specimens which
were pumped out of a gravel pit. Although direct
stratigraphic context is missing, both have a dark
stain imparted byapeatlayer, a taphonomic feature
shared with Pleistocene fauna from the pit that pro-
videsevidencefor temporal association (Frankfort-
er, 1950). The Wet Gravel female is especially dark
and mineralized, while the male is less so, suggest-
ingthatitmightbe more recent. The Prospect burial
was stratigraphically positioned below ash from the
Mount Mazamaeruption around 7000 BP (Cress-
man, 1940). Dates for the remaining crania are primarily
stratigraphic or archaeological and therefore
imprecise, but nonetheless are likely to be early.
Fossil crania were compared to the world database
by Howells (1989) of recent crania. The Howells
samples are mainly from historic populations, and
all are post-Neolithic. Only three of Howells'28 sam-
plesareNative Americans (Arikara, Santa Cruz,
and Peru). We have therefore supplemented the Na-
tiveAmerican samples with six historic samples of
our own: Blackfoot (n 566), Cheyenne (n 522),
Omaha (n 516), Pawnee (n 527), Ponca (n 519),
and Sioux(n528).
Measurements
The general availability of the worldwide cranial
database of Howells (1973,1989) has encouraged
others to record measurements in the same format.
We have followed an expanded version of this pro-
tocolforyears (Key, 1983), producing an extensive
North American database fully compatible with that
ofHowells. Complete measurement sets cannot be
obtained on all crania listed in Table 1. All crania
except Browns Valley were complete or nearly com-
pleteand undistorted. The major reconstruction required
of Browns Valley is attachment of the face,
but the base is unreconstructable. The reconstruc-
tionbyJenks (1937) of the base is unreliable, as he
himself points out. However, Howells'measure-
mentscontain considerable redundancy, and it is
possible to exclude certain variables and still accu-
ratelyquantify morphology. Radii are particularly
useful in this regard, since they parallel standard
measurements taken from basion. Nasion and
bregma radii, for example, are comparable to basion-
nasion length and basion-bregma height, respectively,
except that they are taken from the trans-
meatal axis rather than basion. In addition, radii
quantify lateral facial projections of the upper and
mid-face, which are missed in traditional measure-
mentsets.
The resulting analysis is based on 22 measurements
(Table 2). The designated measurement set
quantifies overall length, breadth, facial variation,
projections from thetransmeatal axis (radii), and
facial projections. Excluded were measurements
TABLE1. Early American crania used in this study
Specimen
Sex
Location
Date
Reference (date; description)
Spirit Cave
M
Nevada
9,415 BP
Dansie, 1997; Jantzand Owsley, 1997
Wizards Beach
M
Nevada
9,225 BP
Dansie, 1997; Owsleyand Jantz, 1999b
Prospect
M
Oregon
7,000 BP
Cressman, 1940
Wet Gravel
F
Nebraska
Frankforter, 1950; Key, 1983
Wet Gravel
M
Nebraska
Frankforter, 1950; Key, 1983
Browns Valley
M
Minnesota
8,900 BP
MysterandO'Connell, 1997; Jenks, 1937
Pelican Rapids
F
Minnesota
7,840 BP
MysterandO'Connell, 1997; Jenks, 1936
Medicine Crow
M
S. Dakota
5,500 BP
Bass, 1976; Key, 1983
Turin
M
Iowa
4,720 BP
Fisher et al., 1985
Lime Creek
M
Nebraska
Key, 1983
Swanson Lake
M
Nebraska
Key, 1983
EARLY AMERICAN CRANIA
147
missing on anyone of the fossil crania, fraction
measurements designed principally to calculate angles,
and difficult measurements possibly subject to
error or generally uninformative (for rationale, see
Jantzand Owsley, 1997).
The specimens in Table 1 were measured in the
same system by three observers. Swanson Lake,
Lime Creek, Medicine Crow, and the Wet Gravel
specimens were measured by Key (1983); Browns
Valley and Pelican Rapids were measured by D.
Hunt; andTurin, Spirit Cave, Wizards Beach, and
Prospect were measured by R. Jantz. All of us
learned Howells'system by reading his excellent
definitions (Howells, 1973) and refining the tech-
niquebytraining on crania that had previously been
measured by Howells. Crania from the Sully site in
South Dakotacomprised the principal training series,
since they were housed at the University of
Tennessee until recently, and constitute one of the
three Native American series used by Howellsinhis
various morphometric studies. Although we have
not conductedaformal interobserver variation analysis,
we are confident that it contributes little to
variation among crania.
Measurements for Spirit Cave, Wizards Beach,
Browns Valley, and Pelican Rapids can be found in
Owsleyand Jantz (1999b). Our measurements for
the other crania are as yet unpublished. The de-
scriptionsofTurin (Fisher et al., 1985) and Medicine
Crow (Bass, 1976) contain some measurements of
those specimens.
Statistical methods
Fossil crania present a number of problems that
make classical statistical approaches impractical.
Each cranium in Table 1 must be considered as
having been drawn from a different population, with
each group represented byasampleofone. It is
therefore not possible to estimate variances and covariances
for use in standard distance and canonical
analyses. Several different approaches have been
employed with fossil material: 1) crania from similar
time periods have been pooled to make a temporally
bounded sample (e.g., Key, 1983; Steeleand Powell,
1992,1994); 2) individual crania have been compared
to recent samples (e.g., Howells, 1995); and 3)
fossil crania have been compared to one another
using a covariance matrix from a large sample of
recent crania (e.g., Van Vark, 1995). The first ap-
proachdoesnot allow examination of variation
among crania. If crania come from populations with
substantial metric differences, pooling yields an unrealistic
average configuration. The second ap-
proachisusefulin examining whether crania resem-
blerecent populations and, if so, which ones. These
comparisons allow historical links between fossils
and extant populations to be hypothesized, but say
little about relationships among fossils. The last
approach allows relationships among fossils to be
examined, which addresses the question of how dif-
ferentiatedthe fossil crania are and whether subpopulations
might be recognized. It requires the as-
sumptionthata covariance matrix obtained from
recent people applies to fossils. This assumption,
although not testable, is reasonable, at least when
applied to Holocene epoch remains.
In this analysis, the second and third approaches
are used. The second approach is well-known in the
classification literature, where an unknown is clas-
sifiedintothe group to which it shows the smallest
Mahalanobis distance. Most statistical package programs
assume that an unknown belongs to one of
the reference populations, and the probability that it
falls into each one of the reference groups is given as
the posterior probability. Mahalanobis D2 has another
important property, namely, to indicate where
a given specimen falls in relation to the variability of
the reference groups. This construct yields what has
been termed the"typicality probability" (Albrecht,
1992). For applications such as the present one, this
is more useful than the posterior probability, because
it is obvious that no early American skull
derives from a contemporary population, regardless
of how similar it maybe.
Mahalanobis D2 between a skull and a sample is
calculated by:
D25~X 2X
j
!9W
21
~X 2X
j
!
where Xis the vector of measurements fora skull, X
j
is the mean vector for populationj, and Wis the
pooled within-sample covariance matrix. D2 can be
referred to a chi-square table withp (number of
variables) degrees of freedom to obtain the typicality
probability (Albrecht, 1992). Mahalanobis D2 between
pairs of fossils can be obtained in the same
way:
D
ij
2
5~X
i
2X
j
!9W
21
~X
i
2X
j
! ,
whereX
i
andX
j
are the measurement vectors for
fossils i and j, and Wis an appropriate covariance
matrix. However, the question of what constitutes
an appropriate covariance matrix is problematic.
Ideally, the covariance matrix should reflect varia-
tionwithingroups with agenetic structure similar
to that of populations from which the fossil crania
were drawn. Inmost cases we do not know what that
population structure was, and it would probably be
difficult to find parallels in recent populations in any
case. The best solution is to takea conservative
approach, using the pooled within-group covariance
TABLE2. List of common measurements
GOL, glabello-occipital length
NAS, nasionsubtense
XCB, maximum cranial breadth
FRC, frontal chord
XFB, maximum frontal breadth
FRS, frontalsubtense
AUB, auricular breadth
PAC, parietal chord
NPH, nasion-prosthionheight
PAS, parietalsubtense
NLH, nasal height
NAR, nasion radius
NLB, nasal breadth
PRR, prosthionradius
OBH, orbit height
FMR, frontomalare radius
OBB, orbit breadth
EKR, ectochonchion radius
DKB, interorbital breadth
ZMR, zygomaxillare radius
FMB, bifrontalbreadth
VRR, vertex radius
148
R.L. JANTZ ANDD.W. OWSLEY
matrix derived from the 34 recent populations used
for comparison.
Defrise-Gussenhoven (1967) showed that the D
between pairs of individuals drawn randomly froma
population will be distributed as= (2p 21) witha
variance of 1, where pis the number of dimensions.
In the present instancep 522, so the random ex-
pectationfortheD between any two crania drawn
from the same population is= (2 z 2221) 56.56.
This random expectation is used to test whether the
distance between pairs of fossil crania is greater
than would be expected if they were drawn froma
single population.
Sexes were pooled by first centering the reference
samples on sex-specific means. Fossil crania were
then expressed as deviations from appropriate sex
means. All computations were performed using software
written by R.L.J.
RESULTS
Relationships among fossils
Table 3presents the matrix of Mahalanobisdis-
tances (D) between each pair of crania. Since the
random expectation is 6.56, any distance greater
than1.65 standard deviations above this value can
be considered significant by a one-tailed test. There
are 55pairwise distances between 11 crania, 14 of
which are significant at the 0.05 level or below. The
significant differences are clearly patterned: 5 of the
14 involvedifferences between Browns Valley and
other crania, and another 4 involve Pelican Rapids
and other crania. The concentration of significant
differences in these two crania marks them as the
most distinctive. The Lime Creek and Swanson
Lake crania account for the remaining significant
differences.
Of greater interest is the general pattern of rela-
tionshipsshownby these crania. A principalcoordi-
nates plot of the distances is shown in Figure 1. The
crania fall into three groups: 1) Browns Valley, Pel-
icanRapids, and Lime Creek are extreme on the
first axis; 2) Turinand Medicine Crow are on the
opposite end of axis one and are separated on the
second axis; and 3) the remaining crania, consisting
of the Wet Gravel specimens, Swanson Lake, Prospect,
Wizards Beach, and Spirit Cave, comprisea
more centrally located cluster. Spirit Cave is somewhat
removed from this cluster in the direction of
the Browns Valley-Pelican Rapids-Lime Creek
group on axis one, but is extreme on axis two. The
Wet Gravelmale departs from the central cluster in
the direction of Turinand Medicine Crow.
These visually defined clusters vary in temporal
and geographic cohesiveness. Medicine Crow and
Turinrepresent Plains Archaic crania. Browns Valley,
Pelican Rapids, and Lime Creek representa
Minnesota-Nebraska group. Temporally this group
is earlier than the Plains Archaic, assuming that the
suspected early date for Lime Creek is confirmed.
The last and largest cluster contains crania from the
Plains, Great Basin, and Northwest. All dated cra-
niainthiscluster are older than 7000 BP.
Relationships to recent groups
Table 4provides the distance and typicality probability
between each fossil skull and the five recent
groups to which it is most similar. They are pre-
sentedinorderof increasing distance from recent
groups. Several points are noteworthy. The first six
fossils (Wet Gravel male, Turin, Wet Gravel female,
Wizards Beach, Prospect, and Medicine Crow) fall
easily within the range of variation of recent groups.
The pattern of similarity to world groups is variable.
The two Plains Archaic crania, Turinand Medicine
Crow, show no particular resemblance to recent Na-
tiveAmericans, and certainly not to those of the
Plains. Medicine Crow has no American Indian sam-
pleamongitsfive most similar groups. Two non-
Plains American Indian groups appear as Turin's
third and fourth most similar groups. The only cra-
niumwithfive American Indian groups as its near-
estneighborsis the Wet Gravel male, while the Wet
Gravel female has none.
The last five crania would be reluctant members of
any recent group. Swanson Lake and Spirit Cave fall
on the margins of some recent distributions, but
Pelican Rapids, Lime Creek, and Browns Valley, all
with extremely low typicality probabilities, would be
highly atypical crania in any recent group used here.
In general, the 11 fossil crania do not show any
particular affinity for the nine Historic period Na-
tiveAmerican samples for which we have data.
TABLE3. Mahalanobis distances (D) between early American fossil crania1
Turin
Prospect
Wizard
Spirit
Pelican
Brown
MedCr
Swan
WGravF
WGravM
Prospect
7.644
Wizard
6.128
5.633
Spirit
7.837
6.401
5.040
Pelican
8.546*
7.956
7.471
6.387
Browns
9.397*
8.514*
8.174
7.342
6.443
MedCrow
6.627
8.013
6.503
8.806*
8.196*
8.525*
Swanson
7.951
8.448*
6.134
6.125
8.463*
8.689*
7.200
WetGravF
7.141
5.942
4.960
5.941
8.463*
8.430*
7.396
5.851
WetGravM
7.247
7.262
6.204
6.360
7.981
7.958
7.702
7.873
6.847
LimeCr
8.977*
7.590
6.779
6.863
6.956
5.529
8.685*
7.811
7.228
8.790*
1
Refer toTable 1 for full names and sex of fossil crania.
* P , 0.05.
EARLY AMERICAN CRANIA
149
These nine American Indian samples represent
26.5%ofthe 34 recent samples used, more than from
any other geographical region. Only two fossils fall
closest to a Native American sample, with the random
expectation being about three. Taking all five
nearest groups for each fossil, 19 of 55 (34.5%) are
American Indians, which is slightly higher than the
random expectation of 15.
Figure 2showsthe principal coordinates plot of
the recent groups with the fossils. This plot was
constructed from the distances among all groups
and the fossils, allowing the fossils to help define the
axes, as recommended by Albrecht (1992). It clearly
shows the separation of Browns Valley, Pelican Rapids,
and Lime Creek from all recent populations on
the first axis. Spirit Cave assumes an intermediate
position. These fossils also have low scores on the
second axis, particularly Lime Creek and Swanson
Lake. This second axis serves to separate recent
Plains populations and Siberia from those of Africa
andAustralasia, as well as from Lime Creek and
Swanson Lake. All other fossils fall closer to or
within the cluster of recent populations.
Browns Valley, Pelican Rapids, Lime Creek, and,
to a lesser extent, Spirit Cave are differentiated on
the first axis by the combination of a wide vault
base, narrow nose, flat frontal bone, and upper facial
forwardness. The Browns Valley cranium was char-
acterizedinmuch the same way by Jenks (1937).
The second axis stresses the narrow vault, short
face, and long parietals of these fossils, particularly
Lime Creekand Swanson Lake, as opposed to the
wide, short vaults and high faces of recent Plains
groups.
DISCUSSION
This analysis includes five crania that are likely
early, but require dating. Until precise dates are
available and additional skeletons studied, it is impossible
to fully assess their meaning. All of them
came into collections prior to the availability of radiometric
techniques. Likely there are additional
skeletons in existing collections that are quite ancient.
Unfortunately, just at the very time when
dating technology using small samples is available,
some institutions and federal agencies are resisting
even minimally invasive chemical or physical analyses.
This restriction impedes understanding of
early American population biology.
In spite of this limitation, our results allow obser-
vationsthat deserve emphasis and have implica-
tionsfortheearly peopling of the New World. The
most significant finding concerns evidence for what
maybe alate Pleistocene/Early Holocene population
Fig. 1. Principal coordinates plot of distances among eleven Paleoand Early Archaic crania from North America.
150
R.L. JANTZ ANDD.W. OWSLEY
that was quite differentiated from other early spec-
imensandfrom recent populations. Evidence for
this population'sexistenceisseen in the Minnesota
specimens and possibly Lime Creek as well. Since
this population existed far from any point of entry
into the New World, it can be argued that it repre-
sentsagroup entering the continent before the an-
cestorsofrecent American Indians. Just how early
depends upon how we model the spread of early
immigrants and upon obtaining additional dates,
especially for Lime Creek. If the assessment by
Howells (1938) of the Torringtoncrania is correct,
an Archaic period population with similar features
persisted in the region until the last millennium
(Agoginoand Galloway, 1963). Whether their differ-
entiationfrom other ancient crania can betaken as
evidence of their descending from different migrants,
or whether they are simply part of an ex-
tremelyvariable early population, cannot be addressed
without better dates and larger samples.
Our identification of the morphological uniqueness
of the Minnesota specimens is at odds with all
post-Hrdlicka assessments of these fossils (e.g.,
Smith, 1976; Owsleyand Jantz, 1999a). The exten-
sivestudybyJenks (1936) of the Pelican Rapids
skull concluded that it possessed a number of prim-
itivefeaturesnot found in recent populations, or
occurred only in low frequency. Hrdlicka (1937a)
responded with an extensive trait by trait comparison,
attempting to show that the Pelican Rapids
skull fell within the range of variation of recent
Sioux. That sentiment has prevailed until recently
in the few post-Hrdlicka assessments of early Amer-
icancrania. Even Jenks (1937) suggested that the
morphology of the Browns Valley cranium was,
apart from certain primitive features, clearly Amer-
icanIndian.
Recent analyses of North American (Steeleand
Powell, 1992,1994; Powelland Rose, 1999) and
South American crania (Neves and Pucciarelli,
1991; Neves et al., 1996) consistently show that
early American crania are differentiated from recent
Native Americans, although these studies have not
indicated the distinct nature of the Minnesota crania.
Like Steeleand Powell (1992), we could argue
that there is a southern Pacific or European similar-
itytosomeofthese crania. Unlike the South Amer-
icansituation, there does not seem to be any partic-
ularresemblance to southwest Pacific or African
populations. The closest example would be the
Swanson Lakecranium, which exhibits features
such as alveolar prognathism, a wide nasal aperture
and interorbital space, guttered nasal sills, anda
short upper facial height, features often associated
with African and southwest Pacific populations.
Swanson Lakeplots closer to these two regional
populations than any other cranium in two-dimen-
sionalspace (Fig. 2), but the overall morphometric
TABLE4. Squared distances (D2) and typicality probabilities of each cranium relative to the five closest modern groups1
Specimens2
West GravelMale
Blackfoot
Arikara
Sioux
Pawnee
Ponca
15.764
21.215
21.237
23.724
25.309
0.827
0.507
0.506
0.362
0.283
Turin
Egypt
Norse
Peru
Santa Cr
Tasman
16.014
16.604
18.905
22.060
27.008
0.825
0.785
0.651
0.456
0.211
West GravelFemale
Hainan
S. Japan
Anyang
N. Japan
Tolai
16.737
19.346
20.668
21.129
21.547
0.778
0.624
0.541
0.513
0.487
Wizard
Norse
Peru
Santa Cruz
Sioux
Blackfoot
17.182
17.346
19.036
19.164
20.041
0.753
0.744
0.643
0.635
0.581
Prospect
Berg
Sioux
Arikara
Pawnee
Ainu
21.259
21.845
22.244
22.308
22.805
0.505
0.469
0.445
0.442
0.413
Medicine Crow
N. Japan
Ainu
Moriori
Philipp
Mokapu
24.316
25.731
25.805
26.385
27.246
0.331
0.263
0.260
0.236
0.202
Swanson
Santa Cruz
S. Japan
Tolai
N. Japan
Ainu
31.089
33.491
34.003
35.519
35.577
0.094
0.055
0.049
0.034
0.034
Spirit
Norse
Blackfoot
Peru
Zalavar
Ainu
32.581
33.718
34.278
34.852
35.836
0.068
0.052
0.046
0.041
0.032
Pelican
Moriori
Norse
Ainu
Pawnee
S. Japan
46.777
47.699
48.038
49.714
49.759
0.002
0.001
0.001
0.001
0.001
Lime Creek
N. Japan
S. Japan
Eskimo
Ainu
Norse
48.740
50.050
52.423
52.617
53.280
0.001
0.001
0.000
0.000
0.000
Browns
Moriori
Mokapu
Arikara
Easter Island
Pawnee
51.767
53.377
54.973
56.814
60.805
0.000
0.000
0.000
0.000
0.000
1
Specimens are listed in the order of increasing distance.
2
For the names and sex of the early American crania (in the left column), refer to Table 1. For the others, see Howells (1973).
EARLY AMERICAN CRANIA
151
configuration includes only the Tolaiof New Britain
as one of its five nearest groups.
It is to be expected that the populations from
which these crania were drawn would be differentiated,
since they are spatially and temporally distinct.
What is surprising is that this differentiation
can be demonstrated with individual crania. Obvi-
ouslysamples consisting of individual crania are not
as powerful as larger samples in demonstrating
variation. That, inturn, suggests that the degree of
differentiation is substantial, although we have not
yet attempted to assess its magnitude. The hetero-
geneityofearly American crania points to the inad-
visabilityof pooling them into a single Paleosample
for purposes of analysis. It is necessary to understand
variation among Paleosamplesin order to
relate cranial variation to models of the peopling of
the Americas.
The heterogeneity observed among the fossil cra-
niainthisstudyis not unique. Van Vark (1994)
demonstrated that European Upper Paleolithiccra-
niaaremore variable than recent Europeans and
even more variable than crania for the entire world.
In Asia, Upper Cave 101 and 103, apparently from
different levels in the cave (Howells, 1983), exhibit
far greater difference than would be expected froma
single population (Cornell, 1998). Whether the high
degree of variation in Europe or Asia is due to frag-
mentedpopulation structure or other processes such
as migration cannot now be specified.
High variability among early American fossil cra-
niamaynotby itself provide evidence of multiple
migrations, but it is consistent with an emerging
consensus that different populations were involved
in the early peopling of North America. MtDNA
evidence, often seen as supporting a single migration
(e.g., Merriweatheretal., 1995; Stone and
Stoneking, 1998), may also be interpreted as supporting
multiple migrations (Schurretal., 1999;
Schurrand Wallace, 1999). MtDNAhaplogroup X,
now recognized as one of the founding New World
haplogroups, suggests ancient connections with Europe
(Brown et al., 1998; Schurrand Wallace, 1999;
Smith et al., 1999), as does lithic point technology
(Stanford and Bradley, 2000). Haplogroup Bis dis-
tributedinAsia and America in away that suggests
it may have arrived via a coastal route.
There is growing realization that the pattern of
cranial morphology seen inmost regions of the world
is relatively recent (Sarich, 1997). The limited avail-
Fig. 2. Canonical plot of world populations and early American fossil crania.
152
R.L. JANTZ ANDD.W. OWSLEY
able evidence suggests that early Asians such as the
Upper Cavespecimens and Liujiangfrom China
(Howells, 1995; Kammingaand Wright, 1988), and
the Gua Gunung specimen from Malaysia (Mat-
sumuraandZuraina, 1999), are not very similar to
recent Asians. Gua Gununghasbeen judged Australian-like
(Matsumuraand Zuraina, 1999), and
recent morphometric analyses of Upper Cave have
argued that they are unlikely ancestors for recent
Chinese (Van Varkand Dijkema, 1988; Kamminga
and Wright, 1988; Cornell, 1998). Recent agricul-
turalexpansions have probably erased much of the
earlier variation present in these regions. The Neolithic
expansions and later migrations in Europe are
well-known and are widely considered to be respon-
sibleforthe recent pattern of genetic variation and
cranial variation (Sokaland Uytterschaut, 1987;
Barbujanietal., 1995). In East Asia, expansion of
rice agriculturists had a similar effect. Van Vark
andDijkema (1988) see Neolithic replacement of the
Upper Cavepeoples as the most reasonable explanation
of the morphometric difference. Schurretal.
(1999) see Neolithic expansion in Siberia as the most
likely explanation of mtDNAhaplotype distribution.
When cranial morphology in the Americas
achieved its modern form is a question that has yet
to be systematically investigated. It has frequently
been observed that the earliest immigrants came
prior to the emergence in Asia of the derived cranial
morphology often termed"Mongoloid" (Angel, 1966;
Lahr, 1995; Neves and Puciarelli, 1991; Soto-Heim,
1994; Steeleand Powell, 1992,1999) In our results,
several crania exhibit metric profiles that fall easily
within the range of variation of recent Native Americans:
Wizards Beach, Prospect, Wet Gravel male,
andTurin. Wizards Beach and Prospect are located
in the West, where the recent form might be ex-
pectedtoappear earliest. The significance of the Wet
Gravel male is difficult to assess in the absence ofa
firm date. Turin, along with Medicine Crow, are the
latest, yet exhibit no close affiliation with any recent
Plains group. This is slender evidence, but argues
that cranial morphology typical of recent tribes in
the Plains had not yet appeared by early to mid
Archaic times.
Other crania, excluded from the present study
because they were too incomplete or data were unavailable,
also illustrate the difference between
Plains Archaic period crania and recent tribes. Lov-
vornetal. (1999) reported on an Archaic burial with
an incomplete cranium from Sidney, Nebraska (C14
date, 4450-4170 BP). They demonstrated that the
cranium is atypical of any historic Plains tribe.
Their analysis using Fordisc 2.0 (Ousleyand Jantz,
1996) places it most similar to Eskimo; it is ex-
tremelylongand high headed. Our own analysis
assigns it to Tasmania, a choice Lovvornetal. (1996)
apparently disallowed. The Lansing cranium, placed
at 5000-6000BPby conventional radiocarbon dates
(Bass, 1973), was pronounced by Hrdlicka (1903) to
show no differences from recent groups in the region.
It is, however, long, narrow, and high, and
atypical of any recent Plains group.
Many regions of North America are known to have
experienced recent incursions. In the Plains, Cad-
doanspeakers expanded northward, displacing the
residentSiouan speakers, and Athapascansmoved
southward. The latter are thought to have hada
large impact on the morphological character of
Plains tribes (Neumann, 1969; Ossenberg, 1994),
seen most dramatically in the lowering of vault
height (Jantzand Willey, 1983). The metric similar-
ityofPlains groups with Siberians (Brennanand
Howells, 1976) provides evidence of recent expan-
sionsoutofAsia. The one Siberian sample in the
present data set, the Buriat, has its lowest distances
with Omaha, Pawnee, Ponca, Arikara, and Sioux,
respectively, rather than with Chinese or Japanese.
Aluinsertions (Novicketal., 1998; see also the mi-
crosatellitedata of Chuetal., 1998) support the
close relationship between recent Amerindians and
East Asians. The most parsimonious explanation of
these morphological and genetic relationships is
that the ancient immigrants have been replaced or
assimilated by more recent ones. This is essentially
the model advocated by Steeleand Powell (1999 and
references therein) after a thorough consideration of
the available evidence.
ACKNOWLEDGMENTS
We thank the following individuals for permission
to study crania in their collections: Amy Dansieand
Donald Tuohyatthe Nevada State Museum; Susan
Mysterand Barbara O'Connellat Hamline University;
and Don Dumondand Guy Tasaatthe Univer-
sityofOregon. Dave Hunt reconstructed the Browns
Valley cranium, and Lee Meadows Jantzproduced
Figures 1and2. Editorial comments were provided
by Sandra Schlachtmeyer, Margaret and Malcom
Richardson, and Roberta Hall. Data collection was
partially funded bya Calhoun Foundation Grant to
the Nevada State Museum and by the National Mu-
seumofNatural History. We thank Roberta Hall for
inviting us to participate in the symposium at which
aversion of this paper was presented in 1998.
LITERATURE CITED
Agogino GA, Galloway E. 1963. The skulls from Torrington, Wyoming:
are-evaluation. Am Antiq 29:106-109.
Albrecht GH. 1992. Assessing the affinities of fossils using canon-
icalvariatesand generalized distances. Hum Evol 7:49-69.
Angel JL. 1966. Early Skeletons from Tranquility, California.
Washington, DC: Smithsonian Contributions to Anthropology,
volume2, no. 1.
BarbujaniG, Sokal RR, Oden NL. 1995. Indo-European origins: a
computer-simulation test of five hypotheses. Am J Phys An-
thropol 96:109-132.
Bass WM. 1973. Lansing man: a half century later. AmJ Phys
Anthropol 38:99-104.
Bass WM. 1976. An early Indian cranium from the Medicine
Crow site, (39BF2), Buffalo County, South Dakota. AmJ Phys
Anthropol 45:695-700.
Brennan MJ, Howells WW. 1976. A craniometricand linguistic
grouping of selected North American and Asian peoples. Unpublished
manuscript.
EARLY AMERICAN CRANIA
153
Brown MD, Hosseini SH, Torroni A, Bandelt HJ, Allen JC, Schurr
TG, Scozzari R, Cruciani F, Wallace DC. 1998. mtDNAhaplo-
groupX: an ancient link between Europe/Western Asia and
North America? AmJ Hum Genet 63:1852-1861.
Cavalli-Sforza LL, Menozzi P, Piazza A. 1994. The history and
geography of human genes. Princeton, NJ: Princeton Univer-
sityPress.
Chu JY, Huang W, Kuang SQ, Wang JM, Xu JJ, Chu ZT, Yang
ZQ, Lin KQ, LiP, WuM, Geng ZC, Tan CC, Du RF, Jin L. 1998.
Genetic relationship of populations in China. Proc Natl Acad
Sci USA95:11763-11768.
Cornell DL. 1998. The morphometric relationship of Upper Cave
101 and103tomodern Homo sapiens . M.A. Thesis, University
of Tennessee, Knoxville.
Cressman LS. 1940. Studies on early man in south central Oregon.
Yrbk Carnegie Inst Washington 39:300-306.
Dansie A. 1997. Early Holocene burials in Nevada: overview of
localities, research and legal issues. Nev Hist Soc Q 40:4-14.
Defrise-Gussenhoven E. 1967. Generalized distance in genetic
studies. Acta Genet Med Gemellol (Roma) 17:275-288.
Fisher AK, Frankforter WD, Tiffany JA, Schermer SJ, Anderson
DC. 1985. Turin: a Middle Archaic burial site in western Iowa.
Plains Anthropol 30:195-218.
Frankforter WD. 1950. The Pleistocene geology of the middle
portion of the Elkhorn River Valley. Lincoln, NE: University of
Nebraska Studies, new series, no. 5.
HaydenblitR. 1996. Dental variation among four prehispanic
Mexican populations. AmJ Phys Anthropol 100:225-246.
Hooton EA. 1930. The Indians of Pecos Pueblo: a study of their
skeletal remains. Department of Archaeology, Phillips Academy,
Andover MA, and New Haven, CT: Yale University Press.
Howells WW. 1938. Crania from Wyoming resembling"Minnesota
Man". Am Antiq 2:318-326.
Howells WW. 1973. Cranial variation in man. Volume 67. Cambridge,
MA: Papers of the Peabody Museum of Archaeology and
Ethnology, Harvard University. 259 p.
Howells WW. 1983. Origins of the Chinese people: interpretations
of the recent evidence. In: Keightley DN, editor. The origins of
Chinese civilization. Berkeley, Los Angeles, and London: Uni-
versityof California Press. p 297-317.
Howells WW. 1989. Skull shapes and the map. Volume 79. Cambridge,
MA: Peabody Museum of Archaeology and Ethnology,
Harvard University. 189 p.
Howells WW. 1995. Who'swhoinskulls. Volume 82. Cambridge,
MA: Peabody Museum of Archaeology and Ethnology, Harvard
University. 108 p.
Hrdlicka A. 1903. The Lansing skeleton. Am Anthropol 5:323-
330.
Hrdlicka A. 1937a. The Minnesota"man". Am J Phys Anthropol
22:175-199.
Hrdlicka A. 1937b. Early man in America: what have the bones to
say? In: MacCurdy GG, editor. Early man as depicted by lead-
ingauthoritiesat the International Symposium at the Acad-
emyofNatural Sciences, Philadelphia, March 1937. London:
Lippencott Co. p 93-104.
Jantz RL, Owsley DW. 1997. Pathology, taphomomyand cranial
morphometricsof the Spirit Cave mummy. Nev Hist Soc Q
40:62-84.
Jantz RL, Willey P. 1983. Temporal and geographic patterning of
relative head height in the Central Plains and Middle Missouri
areas. Plains Anthropol 28:59-67.
Jenks AE. 1936. Pleistocene man in Minnesota. A fossil Homo
sapiens . Minneapolis: University of Minnesota Press.
Jenks AE. 1937. Minnesota'sBrownsValleyMan and associated
burial artifacts. Menasha, WI: Memoirs of the American Anthropological
Association, no. 49.
KammingaJ, Wright RVS. 1988. The Upper Caveat Zhouk-
oudianandthe origins of the Mongoloids. JHumEvol 17:739-
767.
Key PJ. 1983. Craniometric relationships among Plains Indians.
Report of investigations no. 34. Knoxville: Department of Anthropology,
University of Tennessee.
Lahr MM. 1995. Patterns of modern human diversification: implications
for Amerindian origins. Yrbk Phys Anthropol 38:
163-198.
Lovvorn MB, Gill GW, Carlson GF, Bozell JR, Steinacher TL.
1999. Microevolution and the skeletal traits of a middle Archaic
burial: metric and multivariate comparisons to Paleoindians
and modern Amerindians. Am Antiq 64:527-545.
MatsumuraH, Zuraina M. 1999. Metric analysis of an early
Holocene human skeleton from Gua Gunung Runtuh, Malaysia.
Am JPhys Anthropol 109:327-340.
Merriwether DA, Rothhammer F, Ferrell RE. 1995. Distribution
of the four founding lineage haplotypes in Native Americans
suggests a single wave of migration for the New World. AmJ
Phys Anthropol 98:411-430.
Myster SMT, O'Connell B. 1997. Bioarcheologyof Iowa, Wisconsin,
and Minnesota. In: Owsley DW, Rose JC, editors. Bioar-
cheologyofthe North Central United States. Fayetteville, AR:
Arkansas Archeological Survey Research Series. p 147-239.
Neumann GK. 1969. The Siouanpeoplesof the East and the
Plains. Plains Anthropol 14:315-316.
Neves WA, Pucciarelli HM. 1991. Morphological affinities of the
first Americans: an exploratory analysis based on early South
American human remains. JHumEvol 21:261-273.
Neves WA, Meyer D, Pucciarelli HM. 1996. Early skeletal remains
and the peopling of the Americas. Rev Antropol 39:121-
139.
Novick GE, Novick CC, Yunis J, Yunis E, De Mayolo PA, Scheer
WD, Deininger PL, Stoneking M, York DS, Batzer MA, Herrera
RJ. 1998. Polymorphic Aluinsertionsand the Asian origin of
Native American populations. Hum Biol 70:23-39.
Ossenberg NS. 1994. Origins and affinities of the native peoples
of Northwestern North America: the evidence of cranial nonmetric
traits. In: Bonnichsen R, Steele DG, editors. Method and
theory for investigating the peopling of the Americas. Corvallis,
OR: Center for the Study of the First Americans. p 79-115.
Ousley SD. 1995. Relationships between Eskimos, Amerindians,
andAleuts: old data, new perspectives. Hum Biol 67:427-458.
Ousley SD, Jantz RL. 1996. FORDISC 2.0: personal computer
forensic discriminant functions. Knoxville: University of Ten-
nessee.
Owsley DW, Jantz RL. 1999a. Nearsightedness in Paleoamerican
research: historical perspective and contemporary analysis. Paper
presented at Clovisand Beyond Conference, Sante Fe, New
Mexico, October 29-30,1999.
Owsley DW, Jantz RL. 1999b. Databases for Paleoamericanskel-
etalbiology research. In: Bonnischsen R, editor. Who were the
first Americans? Corvallis, OR: Center for the Study of the First
Americans, Oregon State University. p 79-96.
Powell JF, Rose JC. 1999. Report on the osteological assessment
of the"Kennewick Man"skeleton (CENWW97.Kennewick). In:
Report on the non-destructive examination, description, and
analysis of the human remains from Columbia Park, Ken-
newick, Washington[Oct. 1999]. Washington, DC: National
Park Service. htpp://www.cr.nps.gov/aad/kennewick/.
Sarich VM. 1997. Race and language in prehistory. In: Clark GA,
Willermet CM, editors. Conceptual issues in modern human
origins research. New York: Aldinede Gruyter. p 392-410.
Schurr TG, Wallace DC. 1999. mtDNAvariationin native Amer-
icanandSiberians and its implications for the peopling of the
New World. In: Bonnichsen R, editor. Who were the first Americans?
Covallis, OR: Center for the Study of the First Americans,
Oregon State University. p 41-77.
Schurr TG, Sukernik RI, Starikovskaya YB, Wallace DC. 1999.
Mitochondrial DNA variation in Koryaksand Itel'men: population
replacement in the Okhotsk Sea-Bering region during
the Neolithic. AmJ Phys Anthropol 108:1-39.
Smith DG, Malhi RS, Eshleman J, Lorenz JG, Kaestle FA. 1999.
Distribution of the mtDNAhaplogroupX among native North
Americans. AmJ Phys Anthropol 110:271-284.
Smith FH. 1976. The skeletal remains of the earliest Americans:
a survey. Tenn Anthropol 1:116-147.
Sokal RR, Uytterschaut H. 1987. Cranial variation in European
populations: a spatial autocorrelation study at three time periods.
Am JPhys Anthropol 74:21-38.
154
R.L. JANTZ ANDD.W. OWSLEY
Soto-HeimP. 1994. Leshommesde Lagoa Santa (Bresil). Carac-
teresanthropologiques et positionparmid'autres populations
paleoidiennesd'Amerique. Anthropologie 98:81-109.
StanfordD, Bradley B. 2000. The Solutrean solution. Discov
Archaeol2:54-55.
Steele DG, Powell JF. 1992. Peopling of the Americas: paleobiological
evidence. Hum Biol 64:303-336.
SteeleD, Powell JF. 1994. Paleobiological evidence of the peo-
plingofthe Americas: a morphometric view. In: Bonnichsen
R, Steele DG, editors. Method and theory for investigating
the peopling of the Americas. Corvallis, OR: Center for the
Study of the First Americans, Oregon State University. p
141-164.
Steele DG, Powell JF. 1999. Peopling of the Americas: a his-
toricaland comparative perspective. In: Bonnichsen R, editor.
Who were the first Americans? Corvallis, OR: Center for
the Study of the First Americans, Oregon State University.
p97-126.
Stewart TD. 1973. The people of America. New York: Charles
Scribner'sSons.
Stone AC, Stoneking M. 1998. mtDNAanalysisofa prehistoric
Oneotapopulation: implications for the peopling of the New
World. Am JHum Genet 62:1153-1170.
Van VarkGN. 1994. Multivariate analysis: is it useful for homi-
nidstudies? Courier Forsch Inst Senckenberg 171:289-294.
Van VarkGN. 1995. The study of hominid skeletal remains by
means of statistical methods. In: Boaz NT, Wolfe LD, editors.
Biological anthropology: the state of the science. Bend, OR:
International Institute for Human Evolutionary Research. p
71-90.
Van VarkGN, Dijkema J. 1988. Some notes on the origin of the
Chinese people. Homo 39:143-148.
EARLY AMERICAN CRANIA
155