Templeton Human Races-2

Human Races: A Genetic and Evolutionary Perspective
Author(s): Alan R. Templeton

Source: American Anthropologist, New Series, Vol. 100, No. 3 (Sep., 1998), pp. 632-650
Published by: Wiley on behalf of the American Anthropological Association
Stable URL: http://www.jstor.org/stable/682042 .
Accessed: 28/12/2013 09:29
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Wiley and American Anthropological Association are collaborating with JSTOR to digitize, preserve and
extend access to American Anthropologist.
http://www.jstor.org
This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

All use subject to JSTOR Terms and Conditions
ALAN R. TEMPLETON
Departmentof Biology
WashingtonUniversity
St. Louis, MO 63130-4899
HumanRaces:A Geneticand EvolutionaryPerspective

Race is generallyused as a synonym for subspecies,which traditionallyis a geographicallycircumscribed,genetically
differentiatedpopulation.Sometimestraitsshow independentpatternsof geographicalvariationsuch thatsome combinationwill distinguishmost populationsfrom all others.To avoid making "race"the equivalentof a local population,
minimalthresholdsof differentiationareimposed.Human"races"arebelow the thresholdsused in otherspecies, so valid
traditionalsubspeciesdo not exist in humans.A "subspecies"can also be definedas a distinctevolutionarylineage within
a species.Geneticsurveysandtheanalysesof DNA haplotypetreesshow thathuman"races"arenotdistinctlineages, and
thatthis is not due to recentadmixture;human"races"arenot andnever were "pure."Instead,humanevolutionhas been
and is characterizedby many locally differentiatedpopulationscoexisting at any given time, but with sufficientgenetic
contactto make all of humanitya single lineage sharinga common evolutionaryfate. [race, subspecies,lineage, haplotypetree,genetic differentiation]
word raceis rarelyused in the modern,nonhumanevolutionaryliteraturebecauseits meaningis

so ambiguous.When it is used, it is generallyused
as a synonym for subspecies(Futuyma1986:107-109),
but this conceptalso has no precise definition.The traditional meaningof a subspeciesis thatof a geographically
population(Smith
circumscribed,
geneticallydifferentiated
et al. 1997).Theproblemwith thisdefinitionfromanevolutionarygenetic perspectiveis thatmany traitsandtheir
underlyingpolymorphic genes show independentpatternsof geographicalvariation(Futuyma1986:108-109).
As a result, some combinationof characterswill distinguish virtuallyevery populationfrom all others.Thereis
no clearlimitto thenumberof racesthatcanbe recognized
underthis concept, and indeed this notion of subspecies
quickly becomes indistinguishablefrom that of a local
population.One way around this difficulty is to place
minimalquantitativethresholdson the amountof genetic
differentiationthat is requiredto recognize subspecies
(Smithet al. 1997). A second solutionis to allow racesor
subspeciesto be definedonly by the geographicalpatterns
foundforparticular"racial"traitsor characters.A similar
problemis faced in definingspecies. Forexample,thebiological species conceptfocuses attentionon charactersrelatedto reproductiveincompatibilityas thoseimportantin
defininga species. These reproductivetraitshavepriority
in defining a species when in conflict with other traits,
suchas moiphology (Mayr 1970). Unfortunately,thereis
no suchguidanceatthe subspecieslevel, althoughin practice easily observed morphologicaltraits(the very ones
The
deemed not importantunderthe biological species concept) are used. There is no evolutionaryjustificationfor
this dominance of easily observedmorphologicaltraits;
indeed,it merelyarisesfromthesensoryconstraintsof our
own species. Therefore,most evolutionarybiologists reject the notionthattherearespecial"racial"traits.
Because of these difficulties,the modernevolutionary
perspective of a "subspecies"is that of a distinct evolutionary lineage within a species (Shaffer and McKnight
1996) (althoughone shouldnotethatmanycurrentevolutionary biologists completely deny the existence of any
meaningfuldefinitionof subspecies,as arguedoriginally
by Wilson and Brown [1953]-see discussions in Futuyma [1986:108-l O9] and Smith et al. [1997:13]. The
EndangeredSpecies Act requirespreservationof vertebratesubspecies (Pennock and Dimmick 1997), and the
distinctevolutionarylineagedefinitionhas become the de
facto definition of a subspeciesin much of conservation
biology (AmatoandGatesy 1994;Brownlow 1996;Legge
et al. 1996; Miththapalaet al. 1996; Pennock and Dimmick 1997; Vogler 1994). This definitionrequiresthat a
subspecies be geneticallydifferentiateddue to barriersto
genetic exchange that have persistedfor long periods of
time; thatis, the subspeciesmusthave historicalcontinuity in additionto currentgenetic differentiation.It cannot
aloneis
be emphasizedenoughthatgeneticdiff-erentiation
insufficient
todefinea subspecies.The additionalrequirement of historical continuity is particularly important
becausemanytraitsshouldreflectthe commonevolutionaryhistoryof the subspecies,andthereforein theorythere
AmericanAnthropologist100(3):632-650. Copyright(3 1999, AmericanAnthropologicalAssociation

TEMPLETON/
is no need to prioritizethe informativetraitsin defining

subspecies.Indeed,the best traitsfor identifyingsubspecies arenow simply those with the bestphylogeneticresolution.In this regard,advancesin moleculargeneticshave
greatlyaugmentedourability to resolve genetic variation
and provide the best currentresolution of recent evolutionaryhistories(Avise 1994), therebyallowingthe identificationof evolutionarylineages in anobjective,explicit
fashion (Templeton 1994b, 1998a, 1998b;Templetonet
al. 1995).
Thepulpose of this paperis to examinetheexistenceof
races in humans using an evolutionarygenetic perspective. Thefundamentalquestionis: Arehumanpopulations
geneticallydifferentiatedfromone anotherin sucha fashion as to constituteeither sharplygenetically differentiated populationsor distinct evolutionarysublineagesof
humanity?These questions will be answeredwith molecular genetic data and throughthe applicationof the
same, explicit criteriaused for the analysesof nonhuman
organisms.This last point is criticalif the use of the word
race in humanityis to have anygeneralbiologicalvalidity.
This paperwill not addressthe cultural,social, political,
andeconomic aspectsof human"races."
EVOLUTIONARY
GENETICS
OFRACE
abilities.Figure 1 shows thevalues of Fstsandrelatedstatistics for severallarge-bodiedmammals.As can be seen,

the humanFslvalue is one of the lowest, even thoughthe
humangeographicaldistributionis the greatest.A standardcriterionfor a subspeciesorracein the nonhumanliteratureunderthe traditionaldefinitionof a subspeciesas a
geographically circumscribed, sharply differentiated
populationis to have Fstvalues of at least 0.25 to 0.30
(Smithetal. 1997). Hence,asjudgedbythe criterionin the
nonhumanliterature,the humanFstvalue is too small to
have taxonomicsignificanceunderthe traditionalsubspecies definition.
This does not meanthatthe low humanFst
value is without any evolutionarysignificance. Suppose for the moment thatthe Fstvalues in humanstrulyreflect a balance
betweengene flow versuslocal drift/selectionandarenot
due to isolated humanlineages. One convenient method
for quantifyingthis balanceis Nm, the productof local effective populationsize (N) with m, the migrationratebetween demes. Under the idealized population structure
known as the island model, the relationshipbetween Fst
andNmis (Wright1969):
1
Fsr=
Are Human "Races" Geographically

Circumscribed, Sharply Differentiated
Populations?
The validity of the traditionalsubspeciesdefinitionof
humanraces can be addressedby examiningthe patterns
andamountof genetic diversityfound withinand among
humanpopulations.One common methodof quantifying
theamountof withinto amonggeneticdiversityis through
the F5, statistic of Wright (1969) and some of its more
modernvariantsthathave been designed specifically for
molecular data such as Kst(Hudson et al. 1992) or Nst
(Lynchand Crease 1990). Fst
and relatedstatisticsrange
from0 (all the genetic diversitywithina species is shared
equally by all populations with no genetic differences
amongpopulations)to 1 (all the genetic diversitywithina
species is found as fixed differences among populations
with no genetic diversity within populations). The Fst
value of humans(based on 16 populationsfrom Africa,
Europe,Asia, the Americas,and the Australo-Pacificregion) is 0.156 (Barbujaniet al. 1997), therebyindicating
that most human genetic diversity exists as differences
among individuals within populations,and only 15.6%
can be used to genetically differentiatethe majorhuman
"races."To put the humanFstvalue into perspective,humansneed to be comparedto otherspecies. Fsrsfor many
plants, invertebrates,and small-bodied vertebratesare
typicallyfarlargerthanthehumanvalue,butmostof these
organismshave poordispersalabilities,so this is to be expected. A more valid comparisonwould be the Fstvalues
of other large-bodiedmammalswith excellent dispersal
633
4Nm+l
(1)
Most real populationsdo not fit an island model (which

assumes that gene flow is independentof geographical
distance).Nm is thereforenot the actualnumberof individuals exchangedper generation,but ratheris an effective number of migrating individuals per generation
relative to this simple, idealized model of population
structure.This allows comparisonsacross differentspecies in effective amountsof gene flow with respect to a
commonstandard.ForthehumanFstvalueof 0. 156,Nm =
1.35. This resultis consistentwiththe workof Santoset al.
( 1997) who examinedseveralhumandatasets with a variety of statisticalproceduresand always obtainedNm > l .
With Nmon the orderof 1, massive movementsof large
numbersof individualsarenot neededto explainthe level
of genetic differentiationobservedin humans.Moreover,
Nm= 1.35 does not meanthatprecisely 1.35 effective individuals migrate among the "races"every generation;
rather,this is the long-terrnaverage.Assuming a generation time of 20 years,the levels of racialdifferentiationin
humanitycould be explainedby interchangingI .35 effective individualsevely 20 years,or 13.5every 200 years,or
135 every 2,000 years.Since humansoftenmove as populations, gene flow could be very sporadicon a time scale
measuredin thousandsto tens of thousandsof years and
still yield aneffective numberofmigrantsof 1.35.
An Nm value of 1.35 would insurethat the population
evolves as a single evolutionarylineage over long periods
of time (Crow and Kimura1970). Nevertheless,population genetic theoryalso indicatesthatfluctuationsaround


TEMPLETON /
Current
Populations
Asians
Africans
Europeans
EVOLUTIONARYGENETICSOF RACE
Asians
Africans
635
Europeans
1 00,000
Years Ago
Dispersal of
Homo erectus -
l l-
out of Africa
A. AncientOriginCandelabra
B. Recent Origin Candelabra

With Replacement
Figure 2. Candelabramodels of recent human evolution. Part A illustrates the ancient origin version of the candelabramodel. Under this
hypothesis, the major human "races"split from one anotherat the time of dispersal of Homo erectus out of Africa. After that initial split, the
various"races"behavedas separateevolutionarylineages andindependentlyevolved into theirmodernforms.PartB illustratesthe recentorigin
version of the candelabramodel with replacement. Under this hypothesis, an initial candelabraexisted as illustratedin part A. However,
anatomicallymodernhumansthen arose in Africa anddispersedout of Africa around100,000 years ago. This second dispersalevent was marked
by the complete genetic extinction of the earlier Homo erectus populations (indicated by the broken lineages in B) and by a split of these
anatomicallymodernhumansinto separateevolutionaIylineages thatthen independentlyacquiredtheir modern"racial"variation.
theFst
statisticperse cannotdiscriminateamongpotential
causes of genetic differentiation(Templeton 1998a). Althoughhuman"races"do not satisfy the standardquantitative criterionfor being traditionalsubspecies (Smithet
al. 1997), this does not necessarilymeanthatracesdo not
exist in the evolutionarylineage sense. Underthe lineage
concept of subspecies, all thatis needed is sufficientgenetic differentiationto define the separatelineages. If the
lineagessplitonly recently,theoveralllevel of divergence
could be quite small. Therefore,the quantitativelevels of
genetic diversity among humanpopulationsdo not rule
out the possibilitythathuman"races"are valid underthe
evolutionary lineage definition of subspecies. The re-
mainderof this paper will focus upon this more modern

definitionof subspecies.
Are Human "Races" Distinct Evolutionary

Lineages?
Models of Human Evolution and Human Races
When a biological race is defined as a distinct evolutionarylineage within a species, the questionof race can
only be answeredin the contextof the recentevolutionary
historyof the species. The two dominantmodels of recent
humanevolutionduringthe lasthalfof thiscenturyarethe

636
AMERICANANTHROPOLOGIST * VOL. 100, NO. 3
Current
Populations
Africans
Europeans
* SEPTEMBER1998
Asians
Dispersal of
Homo erectzJsout of Afnca
Figure 3. The trellis model of recent human evolution. Under this

hypothesis, Homo erectus dispersed out of Africa and established
populationsin Afnca and southernEurasia,as indicatedby the large
dots. These populations were interconnectedby gene flow so that
there were no evolutionarysublineages of humanityor independent
evolution of the various "races."Arrows with heads on both ends
indicategene flow among contemporaneouspopulations,and arrows
with single heads indicatelines of genetic descent.
candelabra(Figure2) and trellis (Figure3) models. Both

models acceptthe evolutionaryoriginof the genusHomo
in Africa and the spreadof Homoerectusoutof Africa a
million years ago or more. Candelabramodels posit that
the major Old World geographicalgroups (Europeans,
sub-SaharanAfricans,andAsians) split from one another
andsincehavehadnearlyindependentevolutionaryhistories (but perhaps with some subsequent admixture).
Therefore,ffie evolutionaxyrelationshipsamongAfricans,
Europeans,andAsians can be portrayedas an evolutionary tree in this case with the topology of a candelabra
(Figure 2). The major human geographicalpopulations
are portrayedas the brancheson this candelabraand are
therefore valid "races" under the evolutionary lineage
defilnition.The ancientorigin candelabramodel regarded
the split betweenthe major"races"as occurringwith the
spreadof Homoerectus(Figure 2A) followed by independent evolution of each "race"into its modernfortn.
This versionhas been thoroughlydiscreditedand has no
seriousadvocatestoday.However, a recentorigincandelabramodel known as the out-of-Africareplacementhypothesis(Figure2B) has become widely accepted.Under
this model, anatomicallymodernhumansevolved firstin
Africa.Next, a small groupof these anatomicallymoden
humans split off from the African populationand colo-
nized Eurasiaabout 100,000 yearsago, drivingthe Homo

erectuspopulationsto completegeneticextinctioneverywhere (the "replacement"partof the hypothesis).The ancient (Figure2A) andrecent(Figure2B) candelabramodels differonly in theirtemporalplacementof the ancestral
node but sharethe same tree topology thatportraysAfricans, Europeans,and Asians as distinct brancheson an
evolutionarytree.It is thisbranchingtopology thatdefines
"races"underthe evolutionarylineage definition,andnot
the time since the common ancestralpopulation.Hence,
human "races" are valid evolutionary lineages under
eithercandelabramodel.
The trellis model (Figure 3) posits thatHomoerectus
populationsnot only hadthe abilityto move out of Africa
geneticinterchange
butalso backin, resultingin recuITent
amongOld Worldhumanpopulations(LaskerandCrews
1996; Wolpoff and Caspari1997). It is also importantto
note that,underthe trellis model, the taxonomicdesignations of HomoerectusandH. sapiensonly have morphological signifilcanceand do not imply reproductiveisolation as underthe biological-speciesconcept(Mayr 1970).
Therefore,anatomicallymoderntraitscould evolve anywhere in the range of Homoerectus(which includes Africa) and subsequentlyspreadthroughoutall of humanity
by selection and gene flow. Hence, an Africanorigin for
anatomicallymodernhumansis compatiblewith boththe
trellisandcandelabramodels.The two models do differin
their interpretationof interpopulationalgenetic differences. Populationalgenetic differencesreflect the time of
divergencefroma commonancestralpopulationunderthe
candelabramodels. Withthetrellismodel,thegenetic distances reflect the amountof genetic interchangeand not
time of divergence from an ancestor.However, the most
importantdistinction between the candelabraand trellis
models for the discussion at handis thatunderthe trellis
model therewas no separationof humanityintoevolutionary lineages, and hence human"races"are not valid subspecies. In summary,human"races"as evolutionarylineages do exist underthe candelabramodelsbutdo not exist
underthetrellismodel.
Althoughthese two models arefrequentlypresentedas
mutually exclusive alternatives (Wolpoff and Caspari
1997), there is no biological reason why some human
populations may be genetically differentiatedbecause
they arehistoricallineages, whereasotherpopulationsare
differentiatedbecause of recurrentbut restricted gene
flow. Moreover,the genetic differencesbetween any two
humanpopulationsmay representa mixtureof both gene
flow and historical events. Much genetic evidence is
equally compatiblewith bothmodels andhence is noninformative. The emphasis in this paperwill thereforebe
upon data sets that discriminatebetween gene flow and
historicalsplits as non-mutuallyexclusive causes of differentiationamonghumanpopulations.

TEMPLETON/ EVOLUTIONARY
GENETICS
OFRACE
Genetic Diversity Levels within and among

Human Populations
Do the levels of genetic diversity found within and
among human"races"discriminatebetween evolutionary lineage and genetic interchange models of recent
human evolution? As pointed out earlier, F statistics
and related measures of within to among diversity levels do not discriminateper se. However, one conclusion
reached in thatsection has greatrelevance to the debate
over the validity of human races as evolutionary lineages; namely, thatthe estimated gene flow levels in humans are compatible with local differentiation across
geographical space even though the species as a whole
could evolve as a single lineage over time. Much skepticism about the trellis model stems from the belief that a
delicate balance is requiredbetween gene flow (to insure all humans are a common evolutionary lineage
over time) and local genetic drift/selection (to maintain
humans as a polytypic species at any given moment in
time) (Aiello 1997; Nei and Takezaki 1996). Indeed,
even proponents of the trellis model have argued that
only rarely can a species be polytypic under a trellis
model. For example, Wolpoff and Caspari (1997:282)
state that "the humanpattern. . . of a widespreadpolytypic species withmanydifferentecologicalniches . . . is a
very rareone." However, polytypic species are not rare
(Futuyma 1986; Mayr 1970). Moreover, as illustrated
by the examples given earlier, polytypic species occur
over a broadrangeof values forNm andare a robustevolutionaryoutcome. Thereis no difElcultyeither in population genetic theory or observation for the conclusion
thathumanscan be both a polytypic species and a single
evolutionary lineage.
AlthoughF statisticsarecompatiblewith eithermodel
of humanevolution,the claim is made in much of the recent literaturethatwithin "race"diversity levels support
the recent candelabra model. Africans have higher
amountsof genetic diversitythannon-Africansfor many
nuclearloci (Armouret al.1996; Jordeet al.1997; PerezLezaunet al. 1997), mitochondrialDNA (mtDNA) (Comas et al.1997; Francalacciet al.1996), andsome regions
of Y-DNA (Hammeret al.1997). These resultsare often
interpretedas supportingthe recentcandelabramodel by
assumingthatonly a small numberof individualsleft Africa to colonize Eurasiawith little or no subsequentgene
flow. As a result,a bottleneckeffect reducedthe levels of
genetic variationin non-Africans.This interpretationof
genetic diversity also implies that at least Africans and
non-Africansaredistinctevolutionarylineages andhence
are valid races. However, alternativeexplanationsof diversitylevels exist. {Jnderthe neutraltheory,theexpected
heterozygosityfor a DNA region (a standardmeasureof
genetic diversity)is given by:
Heterozygosity=
1+ 4NeF
637
(2)
whereNeis an effective size of the populationand,uis the

mutationrate of the DNA region of interest.Equation(2)
reveals that differences in effective size can explain differencesin the level of genetic diversity.Africansare expected to have higher genetic diversity simply because
theirpopulationsizes were largerduringmuchof the last
million years (Harpendinget al. 1996; Relethford and
Harpending1994, 1995). Indeed, the patternsof genetic
diversityfoundin humansaremoreconsistentwithdifferences in populationsizes and growthratesthanwith differences in population ages from presumedbottlenecks
(Hardinget al. 1997; Perez-Lezaunet al. 1997). The danger of using diversitylevels as an indicatorof population
age froma bottleneckis illustratedby theobservationthat
mitochondrialDNA diversity withinA;fricais higher in
food-producing populations than in hunter-gatherers
(Watson et al. 1996). By equating diversity to age, this
resultwould imply thatagriculturalpeoplesin Africarepresentthe ancestralpopulations,whereasthehunter-gatherers are the recent descendant populations. Such a
conclusion is not credible, andthe diversitylevels within
Africa are interpretedas reflecting effective size differences (Watsonetal. 1996).
Note thatequation(2) has no time component.The reason is that equation (2) describes the diversity levels at
equilibrium.When the equilibriumis disturbedby bottlenecks or rapidpopulationgrowth,time entersas a factor
(Templeton 1997a). Fortunately,differentcauses of departurefromequilibriumcan be discriminated.Forexample, a bottleneck and split should affect all genetic systems. However, nuclear DNA and mitochondrialDNA
show discordantpatternsin humans,a resultinconsistent
with the presumedpopulationbottleneckandthe sharing
by all genetic systems of a commondemographichistory
(Hey 1997; Jordeet al. 1995). One can also discriminate
by thepatternsof diversityacrossgeneticsystemsthatdiffer in mutationrate (Templeton 1997a).A bottleneckreduces genetic variation,and temporaldependenceenters
because mutationtakes time to restoregenetic diversity.
Hence, the longest lasting discrepanciesin relative genetic diversity levels are for low mutationrate systems.
Therefore, under the bottleneck hypothesis, Africans
shouldshow the greatestexcess in relativegenetic diversity for low mutationrate systems. However, the excess
genetic diversityin Africansis foundwith the high mutation ratesystems (Armouret al. 1996;Comaset al. 1997;
Francalacciet al. 1996;Jordeet al. 1997;Perez-Lezaunet
al. 1997), whereas the classic, low mutationratesystems
show comparablelevels of genetic diversity(Bowcock et
al. 1994;Jordeet al. 1995), anda low polymorphicsection
of Y-DNA shows greaterlevels of diversityin Europeans
thanin Africans(Mitchell 1996).An alternativenon-equilibrium pattern can be generated by rapid population

638
* SEPTEMBER1998
growth which causes an increase not a decrease in

levels of genetic diversity.Thehighmutationratesystems
show the earliest and strongest response to increased
populationsize, whichis consistentwiththeobservedpattern.Hence,the observeddiversitypatteinsreflecthuman
populationgrowthratherthanpopulationbottlenecks.
The within "race"genetic diversitylevels do not support the idea that Eurasianssplit off from Africansvia a
small founderpopulation,butthey do not necessarilyfalsify thenotionthata Eurasian/Africansplitoccurredwithout a bottleneck.Therefore,the withinpopulationgenetic
diversitydata are inconclusive on the statusof Eurasians
and Africans as separate evolutionary lineages and
therebyvalidraces.
A.
4J
aJ
u)
Q
o
o
n
Cll
o
C. A. R. Zaire Europeans Chinese Melanesians

Pygmies Pygmies
Genetic Distances and Evolutionary "Trees"

An alternativemethodto Fstof measuringthe extentof
genetic differentiationamong populationsis to convert
the genetic differences into a genetic distance.Thereare
several genetic distance measures available,and sometimes the biological conclusions are stronglydependent
upon the precise measure chosen (Perez-Lezaunet al.
1997).However,thisproblemwill be ignoredin thispaper
because the relative distances among the majorhuman
"races"appearrobustto differinggenetic distancemeasures(Cavalli-Sforza1997). Geneticdistancesin turncan
be convertedinto an evolutionarytree of populationsby
various computer algorithms.Figure 4A shows such a
populationtree(Cavalli-Sforzaet al. 1996).Thisandmost
otherhumangenetic distancetreeshavethedeepestdivergence betweenAfricansandnon-Africans,andthis splitis
commonly estimated to have occurredaround 100,000
years ago (Cavalli-Sforza et al. 1996; Cavalli-Sforza
1997; Nei and Takezaki 1996). All this seems consistent
with the recent candelabramodel, but non-zero genetic
distancescan also ariseandpersistbetweeninterbreeding
populationswith recurrentgene flow (Wright1931, 1943,
1969). As shown by Slatkin (1991), recurrentgene flow
resultsin an averagedivergencetime of gene lineagesbetween populationseven when no population-levelsplit
occurredandthe divergencelevels areat equilibriumand
thereby time invariant.Therefore, an apparentgenetic
time of divergence does not necessarily imply a time of
populationsplitting-or anypopulationsplitat all. Under
a trellis model, genetic distancesreflect the patternsand
amountsof gene flow andnot the age since some "separation"or"split."
Fortunately,these two interpretationsof genetic distancecan be distinguished.If humanpopulationscan truly
be representedas brancheson an evolutionarytree, then
the resultinggenetic distancesshouldsatisfyseveralconstraints.For example, under the candelabramodel, all
non-Africanhumanpopulations'Wsplit"
fromthe Africans
at the same time, and thereforeall genetic distances be-
B.
c. }
Pygl
IA.jeRs
>
3ire
Europeans
Chinese
Za
Imies
Pygl
Melanesians
Figure 4. Genetic distances and recent human evolution. Part A

shows an evolutionarytree of humanpopulationsas estimatedfrom
the genetic distance data given in Bowcock et al. (1991). Human
populationevolution is depictedas a series of splits, and the numbers
on the left indicatethe estimatedtimes of divergence in thousandsof
years. This figure is redrawnfrom Figure 2.4.4, Cavalli-Sforzaet al.
(1996:91). Part B shows the same genetic distance data drawn with
the neighborjoining methodbut without all the constraintsof a tree.
This figure is redrawn from Figure 2.4.5, Cavalli-Sforza et al.
(1996:91).
tween Africanandnon-Africanpopulationshavethe same

expected value (Figure4A). When genetic distances instead reflect the amount of gene flow, "treeness"constraints are no longer applicable.Because gene flow is
commonly restrictedby geographicaldistance (Wright
1943), gene flow models are expected to yield a strong
positive relationshipbetween geographicaldistance and
genetic distance. Figure 4B places the populationson a
two-dimensionalplot in a mannerthatattemptsto reflect
their genetic distances from one another, particularly
nearest-neighbordistances,while otherwiseattemptingto
minimize the total sum of branch lengths (formally, a
neighborjoining dendrogram).Figure 4B uses the same
genetic distance data used to generatethe tree in Figure
4A, but without imposing all the constraintsof treeness
(Cavalli-Sforzaet al. 1996). Note thatEuropeansfall between Africans and Asians as predicted by their geographicallocation in contrastto the candelabramodel
prediction of equal genetic distances of Europeansand
Asians to Africans.Thecomputerprogramsusedto generate "trees"fromgeneticdistancedatawill do so regardless

GENETICS
OFRACE
of what evolutionaryfactorsgeneratedthe distances.It is

thereforethe obligationof the users of such programsto
ensurethatthegeneticdistancedatahave thepropertiesof
treenessbeforerepresentingtheirdataas a tree.To present
trees thatdo not have the propertiesof treenessis analyticallyindefensible,andworse,it is biologicallymisleading.
The failureof humangenetic distancesto fit treenessis
ubiquitous whenever tested (Bowcock et al. 1991;
Cavalli-Sforzaet al. 1996; Nei and Roychoudhury1974,
1982). Nevertheless, these same authorspersist in presentingthe relationshipsof the majorhuman"races"as an
evolutionarytree.Worse,manyrecentpapersdo not even
test for treeness.For example, Nei and Takezaki (1996)
give several trees for both old and new genetic data sets,
butnot a single testof treenessis given or even mentioned.
However, the older data sets given in Nei and Takezaki
(1996) have long been known not to fit treeness(Nei and
Roychoudhury1974, 1982). The newer data sets in Nei
and Takezaki (1996) were not tested for treeness in the
original papers,so a test will be given here using a standard measure of treeness the cophenetic correlation
(Rohlf 1993). Because the trees arethemselvesestimated
fromthegeneticdistancedata,a large,positive cophenetic
correlationis always expected and any correlationless
than0.8 is regardedas a "poor"fit (Rohlf 1993). The copheneticcorrelationsforthenew datasets given in Nei and
Takezaki (1996) are 0.75 for the microsatellitedata of
Bowcock et al. (1994), 0.69 for the microsatellitedataof
Dekaet al. (1995), 0.79 forthe restrictionfragmentlength
dataof MountainandCavalli-Sforza( 1994), and0.45 for
the Alu insertion polymorphism data of Batzer et al.
( 1994). Not one of thedatasets fits treeness.
Inmarkedcontrast,thegeneticdistancedatafit well to a
restrictedgene flow model. In theiranalyses of the older
datasets, Nei andRoychoudhury( 1974, 1982)not only rejected treeness,butshowedthatthe deviationswere those
expected from genetic interchangeamong the "races."
Similarly, Bowcock et al. (1991) not only rejected
treeness for theirdata,but also showed thattheir dataElt
well to a model of"continuous admixture,in time, in
space,or in both:a chainof populationssomewhatsimilar
to a stepping-stonemodel in which the ancestorsof Europeans are geographicallyintermediatebetween the two
extremes,AfricansandAsians"(p. 841). Thephrase"continuous admixture"is an oxymoron, as will be evident
later,but in this case it is used as a synonymfor recurrent
gene flow (Cavalli-Sforza, personal communication).
The "stepping-stonemodel" is a classic isolation by distance model, so Bowcock et al. (1991) show an excellent
Eltof theirdatato the recurrentgene flow model of isolation by distance.Santoset al. (1997) analyzedseveralhuman data sets with a varietyof statisticalproceduresand
found thatthe patternis one of isolation by distancewith
high gene flow between geographically close populations. Finally, Cavalli-Sforzaet al. assembleda compre-
639
0.03
.
av
l)
0.02
0.01
o.oo
1000
2000
3000
4000
5000
Geographic Distance in Miles

Figure 5. Genetic distances and isolation by geographicaldistance.
The global humangenetic distances (the ordinate)are plottedagainst
geographicaldistancein miles (the abscissa).The circles indicatethe
observed values, and the curved line is the theoreticalexpectation
under an isolation-by-distancemodel. This Elgureis redrawnfrom
Figure2.9.2, Cavalli-Sforzaet al. (1996: 123).
hensive humandataset andconcludedthat"theisolationby-distancemodels hold for long distancesas well as for

shortdistances, and for large regions as well as for small
andrelativelyisolated populations"( 1996:124). Figure5
is a redrawingof one of the figures fromCavalli-Sforzaet
al. ( 1996) thatillustrateshow well anisolationby distance
model fits the humandata.
Given that there is no tested humangenetic distance
dataset consistentwith treenessandthatisolationby distance fits the humandata well, proponentsof the recent
candelabramodel have attemptedto salvage the candelabra model by postulatinga complex set of "admixtures
between branchesthathad separateda long time before"
(Cavalli-Sforzaet al. 1996:19). Thekey phrasein thisproposal is betweenbranchesthathadseparateda longtime
before (Terrell and Stewart 1996). Admixture occurs
when genetic interchangeis reestablishedbetweenpopulations that had separatedin the past and undergonegenetic divergence (i.e., the gene flow patternshave been
discontinuous). Proponents of the recent candelabra
model then attemptto reconcile the genetic distancedata
with an admixturemodel thatmimics some of the effects
(andthe good fit) of recurrentgene flow. By invokingadmixture events as needed, human "races"can still be
treatedas separateevolutionarylineages,butnow withthe
qualificationthatthe "races"were purerin the past the
paradigmof the "primitiveisolate" (Terrelland Stewart
1996). However, even advocatesof the recentcandelabra
model acknowledge that these postulated admixture
events are "extremelyspecific" and"unrealistic"(Bowcocketal. 1991:841).

640
1998
* VOL. 100, NO. 3 * SEPTEMBER
ANTHROPOLOGIST
AMERICAN
For example, consider Melanesiansand Africans. As

shown in Figure 4B, these two humanpopulationshave
nearlymaximalgenetic divergencewithin humanityas a
whole with respectto molecularmarkers.Moreover,note
thatEuropeansarecloser to bothAfricansandto Melanesians thanareAfricansto Melanesians(Figure4B). However, MelanesiansandAfricanssharedarkskin, hairtexture,andcranial-facialmolphology (Cavalli-Sforzaet al.
1996; Nei andRoychoudhury1993) the traitstypically
used to classify peopleintoraces.Oneobvious conclusion
from this gross disparitybetween racially defining traits
andthemoleculargeneticdatais thatclassificationsbased
on these "racial"traits have no evolutionary validity.
However, in orderto salvage the racial types emerging
from the candelabra model, Nei and Roychoudhury
(1993) propose two dispersal events out of Africa. The
first group of people moved throughthe Middle East to
Northeast Asia and then moved southward to occupy
SoutheastAsia. Later,a secondgroupof humansmigrated
out of Africato theIndiansubcontinentandthento Southeast Asia, where admixture occurred with the earlier
Asian group.Nei andRoychoudhurythenproposethatthe
resultantadmixedpopulationin SoutheastAsia absorbed
most of its gene pool from the older Asian group,but"retainedthe genes fordarkskin,frizzledhair,etc. fromAfricans, becauseof naturalselection in tropicalconditions"
( 1993:937).This admixedgroupthenmoved out to the islandsof thePacificandAustralia.Thepartof this admixed
populationthatremainedin SoutheastAsia andIndiathen
experienced additionaladmixtureevents involving the
olderAsiansandEuropeans.This secondroundof admixture wiped out most of the "Africantraits"in India and
SoutheastAsia except for a few isolated subpopulations
(Nei andRoychoudhury1993).
Nei andRoychoudhuryarguethatthis complicated,ad
hoc scheme is more plausible than the hypothesis of
"independentevolution of African traits in this area"
(1993: 938). However, no mention is even given to the
trellis model interpretationin which these traits are not
"African"traitsat all, but rathertropical adaptive traits
thatarefavoredin humanpopulationsliving in the appropriate environment-populations that are not evolutionarily independentbecause they were and are in genetic
contact.Moreover,even this complicatedscheme of multiple admixtureand massive populationmovements still
does not explainthe geneticdistancedata.Admixedpopulations areexpectedto be intermediatein genetic distance
between the original parentalpopulations,but Melanesians arenot intermediatebetweenmainlandAsian populationsandAfricans(Figure4B). This exampleshows that
althoughcomplex, multiple ad hoc admixtureevents are
invokedto reconcilethe recentcandelabramodel with the
genetic distancedata, they still fail to do so. In contrast,
isolationby distancefits the humandatawell andall thatit
requiresis thathumanstendto mateprimarilywith others
bornnearbybut often outsideone's own natalgroup

(LaskerandCrews1996;Santoset al. 1997).
The hypothesisof admixturecan be testeddirectly.
occursbetweenbranchesthathavedifWhenadmixture
underpastisolation,geneticclinesareset up
ferentiated
loci.Thisresultsin a
forall differentiated
simultaneously
in theclinesforall geconcordance
stronggeographical
neticsystems,bothneutralandselected.Incontrast,isolaconcordance
tionby distancemayresultin geographical
forsystemsundersimilarselectiveregimes(Endler1977;
is
KingandLawson1997),butotherwisenoconcordance
expected.Hence,the lack of concordanceof "African
traits"withmoleculargeneticdistancesis notsurprising
underanisolationby distancemodel.Thelackof concorof differenteledancein the geographicaldistribution
mentshasbeenthoroughlyandextensivelydocumented
argutraditional
byothersandhasbeenoneof theprimary
ments againstthe biologicalvalidityof humanraces
et al. 1996;Futuyma1986).Thislackof
(Cavalli-Sforza
acrossgeneticsystemsfalsifiesthehypotheconcordance
sis of admixtureof previouslyisolatedbranchesandthe
inthepast.
were"pure"
ideathat"races"
The geneticdistancedata are thereforeinformative
aboutthe statusof human"races"as evolutionarylineages. Geneticdistanceanalysesstronglyanduniformly
indicatethat human"races"cannotbe representedas
treeas underthecandelabra
brancheson anevolutionary
events.Gemodels,evenby invokingadhoc admixture
the
neticdistances,whenproperlyanalyzed,undermine
linebiologicalvalidityof humanracesas evolutionary
ages.
HaplotypeTrees
Thefinaltypeof geneticevidenceto be consideredis
of thegethatarisingfromphylogeneticreconstructions
neticvariationfoundin homologousregionsof DNAthat
All thehomologouscopshowlittleornorecombination.
ies of DNA in such a DNA regionthatareidenticalat
everynucleotide(orinpractice,identicalatallscorednucleotidesites)constitutea singlehaplotypeclass.A mutation at anysite in thisDNA regionwill usuallycreatea
new haplotypethat differs initiallyfrom its ancestral
change.As timeprohaplotypeby thatsinglemutational
ceeds,somehaplotypescan acquiremultiplemutational
changesfromtheirancestraltype.All thedifferentcopies
of a haplotypeforeachof thehaplotypesin a speciesare
subjectto mutation,resultingin a diversityof haplotypes
closenessto
in thegenepoolthatvaryin theirmutational
in the
one another.If thereis littleor no recombination
DNA
DNAregion(asis thecaseforhumanmitochondrial
orforsmallsegmentsof nuclearDNA),thedivergenceof
haplotypesfromone anotherreflectsthe orderin which
history.Whenmutamutationsoccurredin evolutionary
tional accumulationreflectsevolutionaryhistory,it is

GENETICS
OFRACE
possible to estimatea networkthatshows how mutational

changes transformone haplotype into another or from
some commonancestralhaplotype.Sucha networkrepresents an unrootedevolutionarytreeof the haplotypevariation in thatDNA regionandis called a haplotypetree.In
some circumstances,the ancestralhaplotypeis known or
can be inferred,therebyprovidinga rootedhaplotypetree.
Inpractice,haplotypetreesaresometimesdifficultto infer
from the mutationaldifferencesamong a set of observed
haplotypesbecausethe samemutationmayhave occurred
more than once, therebydestroyingthe relationshipbetween mutationalstate and evolutionaryhistory, and/or
recombinationmayhave scrambledupthe DNA regionso
thoroughly that accumulatedmutationaldifferences reflect both evolutionaryhistory and recombinationin a
confoundedfashion.Whenthey can be estimated,haplotype trees directlyreflectonly the evolutionaryhistoryof
the genetic diversitybeing monitoredin the DNA region
under study. Haplotypetrees are not necessarily evolutionarytreesof species norof subpopulationswithin species. For example, suppose a species is and always has
been completely randomlymatingas a single population
andthereforehasno subpopulationevolutionaryhistoryat
all; yet that same randomly mating species will have
haplotype trees for all homologous DNA regions that
show little orno recombination.
The publication of mitochondrial haplotype trees
(Cannet al. 1987;Vigilantet al. 1991) motivatedmuchof
thecurrentdebateoverrecenthumanevolution.These and
subsequent papers (Stoneking 1997) on mitochondrial
haplotypetrees makea threefoldargumentin favor of the
recent candelabramodel: (1) all mtDNA types in current
humanpopulationscan be tracedbackto a single common
ancestor(mitochondrial"Eve"),(2) the root of the mitochondrialtree is in Africa,and (3) the treecoalesces to its
common ancestral type about 200,000 years ago. Althoughthe originalhaplotypetreeswere estimatedincorrectly because of an improperuse of a computerprogram
(Maddison 1991;Templeton1992), this erroris trivialin
light of thefact thatthesethreeargumentsarenoninformative aboutthe statusof humanpopulationsas evolutionary
lineages and thereforedo not discriminatebetween the
candelabraand trellis models (Templeton 1994a). Point
( 1) is a universalforall modelsof human(andindeed,nonhuman) evolution because all homologous segments of
DNA are expected to coalesce to a common ancestral
molecule underany model of evolutionin a finite population (Tavareet al. 1997). Indeed,haplotypetrees would
not exist at all if this were not true.With respect to point
(2), the trellis model is compatiblewith any root location
occupied by humansat the time of coalescence, which includes Africa. Because the bulk of humanitylived in Africa hundredsof thousandsof years ago (as previously
noted), an Africanroot is the most likely result underthe
trellis model. Argument(3) is based on the premise that
641
mitochondrialDNA can spread only when populations

expand geographically, so mitochondrial DNA either
spreadwithHomoerectus(a million yearsago or more)or
with the presumed spread of anatomicallymodern humansabout 100,000 years ago (Cannet al.1987; Stoneking 1997; Vigilant et al. 1991). This premiseequatesthe
mitochondrialhaplotypetree to a populationtree.Haplotype trees may or may not reflect populationhistory(indeed, as pointed out above, there may be no population
historyatall), andthispropositionneedsto be testedrather
than assumed. In particular,when dealing with populations thatareexchanginggenes (the premiseof the trellis
model),a haplotypecan spreadgeographicallyatanytime
via gene flow. Hence,a coalescence timeof 200,000 years
ago is compatiblewith either model of humanevolution
(Templeton1994a).
A fourthargument,not presentin theoriginal"Eve"papersbutrelatedto mtDNA coalescencetime,is thatthehuman populationsize at the time of coalescence was too
small to be compatible with the trellis model (Rogers
1997). Underneutrality,the expectedcoalescencetime of
mtDNA is 2Ne
generations,whereNeisthe inbreedingeffective size of females. Assuming a coalescence time of
200,000 years ago and a generationlength of 20 years
yields Ne=5,000. More complicatedcoalescent models
yield differentestimates,but all are on the orderof thousandsforNe(Rogers 1997). Neisnot the census size offemales. In general, effective sizes are much smaller than
census size. For example, in conservationbiology it is
standardto assumethatthe effective size is only one-fifth
the census size for large-bodiedmammals.This fivefold
correctionfactor from conservation biology assumes a
stableor decliningcensus size, but whenpopulationsizes
areincreasing,as seems to be the case forhumansover the
past hundredthousandyears or so, inbreedingeffective
size canbe ordersof magnitudesmallerthancensussize or
othereffective sizes, such as the varianceeffective size
(Templeton 1980). Hence, a fivefold correctionfor inbreedingeffective size to census size is undoubtedlyconservativefor recenthumanevolution. Moreover,the census size should be doubled to include males. Thus, the
estimateof Ne= 5,000 implies a census size of 50,000 humansor more. Also, coalescence time is not knownto be
exactly 200,000 years but ratherhas a broadconfidence
intervaldue to a lack of precise knowledgeaboutthe neutralmutationrate and evolutionarystochasticity(Tavare
et al.1997; Templeton 1993). Using the full rangeof ambiguitygiven in Templeton(1993), populationsizes up to
200,000 cannot be excluded. Moreover,since 1993, the
ambiguityon the mtDNA mutationrate has actually increased(Arnasonet al.1996; Howell et al.1996; Parsons
etal.1997), takingthe upperlimitsofthe confidencerange
close to a populationsize of 500,000. All of these calculations dependuponthe assumptionof neutrality.Deleterious mutationswill cause this procedureto underestimate

642
1998
* SEPTEMBER
effective size, andsuchmutationsareknownto exist (Hey

1997;Nachmanet al. 1996;Templeton1996). Therefore,
all of the above calculationsare lowerboundsgiven this
demonstratedviolation of assumptions. More importantly, even a singleadvantageousmutation occurring
within the mtDNA genome at anytimeduring
anywhere
thepastfew hundredsof thousandsof yearsof humanevolutionwill maketheeffective size estimatorquantitatively
meaningless(Rogers 1997). Given the broadconfidence
ranges associatedwith this estimationprocedureand its
extraordinarysensitivity to deviationsfrom neutrality,it
is patentthatthe populationsize argumentdoes not discriminateamongthe alternatives.
Fortunately,there is much informationin haplotype
trees that can be used to test the hypothesis that human
"races"are evolutionary sublineages whose past purity
has been somewhatdiminishedby admixture.Forexample, in orderto reconcilethecandelabramodelwiththegeneticdistancedata,it is necessaryto regardEuropeansas a
heavily admixed population (Bowcock et al. 1991;
Cavalli-Sforza et al. 1996). When admixture occurs,
haplotypesthatdiffer by multiplemutationalevents with
no existing intermediatehaplotypesshouldcoexist in the
admixedpopulation'sgene pool (ManderscheidandRogers 1996;Templetonet al. 1995). The detectionof such
highlydivergenthaplotypesrequireslargesamplesizes of
thepresumedadmixedpopulationin orderto havestatistical power. When large sample surveys have been performed upon the presumed admixed Europeanpopulations, no highly divergent haplotypes or evidence for
admixtureareobservedfor eithermtDNA (Manderscheid
andRogers 1996) or Y-DNA (Cooperet al. 1996).Incontrast, isolation by distance (the trellis model) produces
gene pools without strongly divergent haplotypes (i.e.,
most haplotypesdiffer by one or at mostjust a few mutationalsteps from some otherhaplotypefoundin the same
population),as is observed.
The candelabraand trellis hypotheses are models of
how genes spreadacross geographicalspace andthrough
time,andhencea geographicalanalysisof haplotypetrees
providesa directtest of these two models. Statisticaltechniques exist that separate the influences of historical
events (suchas populationrangeexpansions)fromrecurrentevents (such as gene flow with isolation by distance)
whenthereis adequatesamplingbothin termsof numbers
of individualsand of numbersand distributionof sampling sites (Templeton et al. 1995). This statisticalapproachfirstconvertsthe haplotypetreeintoa nestedstatisticaldesign.The lowest level of analysisis the haplotypes
themselves, and the Elrstlevel of nesting is created by
startingat the tips of the haplotypenetworkand moving
one mutationalstep in, forminga unionof any haplotypes
that are reachedby such a single mutationalstep or that
converge upon a common node. This first set of "l-step
clades" (Templetonet al. 1987) on the tips of the haplo-
type networkis then prunedoff and the process repeated

untilall haplotypesareincludedin 1-stepclades.Now one
has a tree of 1-stepclades, andthis treecan be nested into
"2-step clades" using exactly the same nesting rules, but
using 1-step clades insteadof haplotypesas the base unit.
These nesting rules are used at successively higherlevels
untilthe next level of nestingwouldplace the entireoriginal haplotypetreeinto a single clade (formoredetails,see
TempletonandSing 1993).
The age of a higherordernestingclade has to be as old
or olderthanthe clades nestedwithinit. Thus,even in the
absence of a root for the haplotypetree,the nested design
providesrelativeage information.By studyinghow a series of nested clades is distributedin space, it is therefore
possible to makeinferencesabouthow haplotypelineages
spreadgeographicallythroughtime. Moreover,the geographical range of a clade relative to that of the other
clades it is nested with at the next higher level indicates
how farspatiallya haplotypelineagecan spreadduringthe
time it takes to accumulatea single mutation.Hence, the
nested design based on the haplotypetree automatically
adds a temporaldimension to the spatial data gathered
with the sampleof currenthaplotypes.It is thereforepossible to reconstructthe historicaldynamics of the geographicalspreadof haplotypelineages, with the dynamical resolutionbeing limitedby theaverageamountof time
it takes a lineage to accumulatea single mutation.Moreover, by makingthe analysisnested,no assumptionof homogeneity is being madeabouthow lineages spreadgeographically over time. That is, at one time or place,
haplotypelineages may havespreadthroughgene flow restrictedby geographicaldistance;atanothertime orplace,
there may have been a rapidrangeexpansion;and at yet
anothertime orplace, all geneticinterchangebetweentwo
geographicalregions may have been severed.The nested
analysis does not exclude any of these possibilities a priori,butratherregardsall of them(oranymixture)as legitimatefactorsinfluencingthe movementof haplotypelineages throughtime andspace(Templetonet al. 1995). This
statistical approachthereforetreatshistorical and recurrentevents asjointpossibilitiesratherthanas mutuallyexclusive alternatives.
These differentfactors,however,leave differentsignaturesin the nestedanalyses.If gene flow restrictedby isolation by distance dominatedduringthe place and time
when a certainsubsetof mutationsoccurred,thenthe older
clades defined by these mutationsshould be more widespread and the younger but evolutionarilyclose clades
shouldbe in the same generalareaas theolderclades. This
expectationfollows fromthe simplefact thatunderisolation by distance,genes spreadonly a little every generation, andthelongera gene lineageexists, themoregenerations it has to spreadgeographicallyand to accumulate
additional mutations. If two geographicalregions split
from one another(i.e., severedgenetic interchange),then

TEMPLETON /
the clades that markthose geographicalregions and that

time of isolationwould accumulatemanymutationaldifferences but without movementinto each other's space.
Finallyeif a subset of the originalpopulation(containing
only a subsetof the haplotypevariationthatexistedat that
time) suddenlyexpandedinto and colonized a new geographicalregion, then the subset of haplotypesthey carriedandthelineagesderivedfromthemwouldhave widespread geographical distributions for their frequency
relativeto the populationas a whole. Thus,gene flow and
differenthistorical events leave distinct genetic-spatial
signaturesin a nestedanalysisandaretherebydistinguishable. Moreover,the areas affected by these forces and
events can be inferred, as can their time relative to the
nesteddesignof the haplotypetree.
The ability to discriminatethe genetic signaturesof
rangeexpansionsfrom recurrentbutrestrictedgene flow
is criticalto discriminatingthe candelabrafromthe trellis
models andtherebyinferringthe evolutionaryvalidityof
race.Thecriteriaused to identifyrangeexpansionsin this
nestedapproachhave been empiricallyvalidatedby analyzingdatasets with strongpriorevidenceof rangeexpansion andwere found to be accurateandnot proneto false
positives (Templeton1998a).Applicationof this statistical approachto humanmtDNA haplotypetreesyields the
643
significant results summarizedin Figure 6 (Templeton

1993,1997b,1998a).
As shown in Figure6, humanmtDNA yields a pattern
of isolation by distancebetweenAfricans and Eurasians
throughoutthe entire timeperiodmarkedby mtDNAcoalescence (Templeton 1993,1997b), therebysignificantly
rejectingboth the candelabrahypothesisof no gene flow
between Africans and non-Africansand the admixture
models used to reconcilethe candelabramodels with the
genetic distancedata.Recurrentgene flow in this analysis
is relativeto the time scaledefinedby the coalescence and
mutationratesof mtDNA,so gene flow amongOld World
humanpopulationscould have been sporadic on a time
scale of severaltens of thousandsof years.
Figure6 also revealsthatrangeexpansionsplayeda signiElcantrole in recenthumanevolution.Among the statistically significantrangeexpansionsis a relatively recent
rangeexpansionacrossEurope(Templeton1993,1997b),
an inference supportedby other mtDNA data sets (Calafell et al. 1996; Comas et al. 1997; Francalacciet al.
1996). A recent study on mtDNA isolated from a Neandertal(Kringset al. 1997) is suggestive (but not conclusive as the sample size is one) that Neandertalswere replaced in Europe.This inferenceis compatiblewith the
statisticallysignificantEuropeanexpansionshownin Figure6, butfurtherdataareobviouslyneededto determineif
Restricted Gene Flow After

Range ExpansionMostWThrough
Isolation by Distance But With
<
Some RecurrentLong
<
Distance Interchange
_
a
R
f
_,^
Figure 6. Statisticallysignificant inferencesfrom geographicalanalyses of human mtDNA haplotypetrees. As far back as is observable with
mtDNA, there was gene flow restrictedby isolation by distance in human populations living in Africa and southernEurasia. Mo-rerecent
statistica}lysignificantrange expansionevents are indicatedby wide arrows.There were expansions into Europe,northernAsia, the Pacific, and
the Americas.Two arrows are indicatedgoing into North America because this expansion either involved a colonization event with a large
numberof people, an extendedcolonization,or at least two separatecolonizationevents. The lines drawnthroughthese arrowsindicatethatafter
the colonizationthere was a significantreduction,perhapscessation, of gene flow between Asia and NorthAmerica.After the colonization of
North America,there were furthersignificantexpansionsinto the remainderof the Americas.After these expansionevents, there is statistically
significantgene flow once again.Most of this postexpansiongene flow fits the expectationsof isolationby distance,but some postexpansiongene
flow occurredthroughlong-distanceinterchanges.

644
* SEPTEMBER1998
this recentEuropeanexpansionevent was also a replacement event. The other recent expansions (into northern
Asia, the Pacific,andthe Americas)appearto be rangeexpansionsintopreviouslyunoccupiedareas.
Genetic interchangebetween Afficans and Eurasians
over long periods of humanevolutionaryhistory is also
stronglysuggestedby a hemoglobinbetalocus tree(Harding et al. 1997). The coalescence of an autosomalgene is
expectedto be aboutfourtimes as old as thatof mtDNAor
Y-DNA, and this seems to be the case for the beta locus
(Hardinget al. 1997). Consequently,the patternsof widespreadgene flow acrossAfricaandAsia observedwiththe
hemoglobin locus predate the hypothesized "replacement"eventof therecentcandelabramodel(Hardinget al.
1997). Obviously, if such a replacementhad occurred,
these earliergenetic signaturesof gene flow shouldhave
beenobliterated.
To reinforcetheseconclusions,thehemoglobinbetalocus dataof Hardinget al. ( 1997) were subjectedto a nested
clade analysisof geographicalassociations(Templetonet
al. 1995). Firstthe estimatedhaplotypetree is converted
into a series of nestedbranches(clades) (Templetonet al.
1987; Templetonand Sing 1993). Figure7 shows the hemoglobin haplotype network of Harding et al. (1997),
along with the nested statisticaldesign. Once the haplotype tree has been converted into a nested statisticalde-
sign, the geographical data are quantifiedin two main

fashiorls(Templetonet al. 1995): the clade distance,Dcs
which measures the geographicalrange of a particular
clade; and the nested clade distance,Dnwhich measures
how a particularclade is geographicallydistributedrelative to its closest evolutionarysisterclades (i.e., clades in
the same next higher-levelnestingcategory).Contrastsin
these distance measures between older and younger
clades areimportantin discriminatingthepotentialcauses
of geographicalstructuringof thegenetic variation(Templetonetal. 1995), as discussedabove.Inthiscase, temporal polarityis determinedby an outgroupanalysisthatindicates thathaplotypeB3 in Figure7 is the root (Harding
et al. 1997)in additionto thepolarityinherentin thenested
design itself. The statisticalsignificance of the different
distancemeasuresand the old-youngcontrastsare determined by randompermutationtesting that simulates the
null hypothesisof a randomgeographicaldistributionfor
all clades within a nesting category given the marginal
clade frequenciesand sample sizes per locality. Figure 8
presents the results of this nested clade analysis of geographicaldistributions.
The statisticallysignificantpatternsshown in Figure 8
need to be interpretedbiologically.In orderto makeinference explicitandconsistent,a detailedinferencekey is providedas an appendixto Templetonet al. (1995) (hereafter
Figure 7. Ee hemoglobinhaplotypenetworkof Hardinget al. ( 1997), alone with the nested statisticaldesign. Haplotypedesignationsare those
given in Hardinget al. ( 1997). Nested groupingsabove the haplotypelevel are designatedby "C-N,"where C is the nesting level of the clade and
N is the numberof a particularclade at a given nesting level. Boxes with thin lines nest togetherhaplotypesinto l-step clades, and boxes with
thick lines nest together l-step clades with 2-step clades.

OFRACE
GENETICS
Haplotypes
ll
l-Step Clades
Clade
Dc
Dn
1-1
6532
6519
2-Step Clades
Clade
Dc
.,,,-.--''...,t,...,g'. ,.,,6-..:'.-
.. ;"..,,--.,.-.'
:..:
Dn
Dc
|| Clade
O-Y
6469
-2646
B1
6308
6324
108
-1579
O-Y
'; iSA
-m' -"'S-0 '' ' w
'-
0' ' 1 " ' ' i ' l " ' 1 '' " 0
"
igE'
4867
1586
B9
_
"'
- ' ' '-''',-,'* -Ba'-; -,"""
_
7039
A4
_
6542
A3
_
13784
A2
1-2-11-1' No: RE
j::-;iL
jititi-100i;i50
;0-:t;
::itE;j--0000000-;i:;2ii2:0-;-t0;t:
:::f:;;:it:B2:
:::
iT'-E
Tr.
:: .E
h?
E f ::
7 ) i
i:E: 0 i-
iDE
i iS i:iEi :.
:E iE-E :E i E. .E.E i,
i- iR -E i! iR iEE ..
i.
B4
3054
Bll
3351
O-Y
491
1640
1-2-34 No: IBD
i 8titi
0;0
ii -iiii4)
;''-ti'' 'tEg'
..,,X,...'.-. SS'''.':-..',.,,.'--"
3745Ss 6515SS
C7
_
gO98
8478
C2
glll
lOSlllL
Cl
?-Y
681
2407
O-Y
4632S
9626SS
1-3
_
-6916SS
E i...i.
E.fi- :ET.EE
i-
:::7E
iS
ti-ii00i
-i
...E100.::i:iE.,!--iCifi.
Li::40
;:-ES
:E
22L
E
1-^-11-1^-13-14No:
7
D1
5795
5855
O-Y
-S725
3940
|
O-Y
|
6209
2402
2-2
81861L 874811
? Y
-2088SS
_2547SS
1-2-11-12No. RE
6394S
2072L
1-2-34 No: IBD
Figure 8. Resultsof the nestedgeographicanalysisof thehumanbeta

chainhemoglobinhaplotypes.The nested design is given in Figure7,
as are the haplotypeand clade designations.Following the name or
numberof any given clade are the clade and nested clade distances.
The oldest clade within a nested group is indicatedby shading.The
average difference between the oldest and younger clades within a
nesting categoIy (as determined by B3 being the root) for both
distance measures is given in the row below a dashed line labeled
"O-Y."A superscriptS means that the distance measureis signiE1cantly small at the 5% level, and SS, at the 1% level. Similarly, a
superscrsptL means thatthe distancemeasureis signiElcantlylarge at
the 5% level, andLL, at the 1%level. At the bottomof the boxes that
indicatea nested set of clades in which one or more of the distance
measuresis significantlylarge or small is a line indicatingthe biological inference.The numbersrefer to the sequenceof questions in
the TRP key thatthe patterngenerated,followed by the answerto the
filnalquestionin the TRP key. Following this answeris the biological
inference generated by use of the TRP key, where RE is range
expansion and IBD is recurrentgene flow restnctedby isolation by
distance.
referredto as the TRP key). Templeton(1998a) gives an

empiricalvalidationof this key. This key providesfor the
of the distinctsigobjectiveandsystematicidentiE1cation
naturesassociatedwith isolationby distance,fragmentation, and range expansion that were describedqualitatively above. Moreover, the key also identifies the
artifactsthat can emerge from inadequategeographical
sampling.Consequently,not all rejectionsof the null hypotheses can be interpretedbiologically. Figure8 shows
theresultinginferences.
In comparingthe mt9NA (Figure6) and hemoglobinL
(Figure8) inferences,it is importantto keeptwo factorsinU
645
mind.First,thesetwohaplotypetreesaredetectingevents
thetimedepthof the
ondifferenttimescales.Inparticular,
hemoglobinnetworkhas a 95Soconfidenceintervalof
400,000 to 1,300,000yearsago (Hardinget al. 1997).
Once ultimatecoalescencehas occulTedin a haplotype
aboutpreviouseventsorevotree,thereis noinformation
lutionaryforces.Therefore,the oldereventsandforces
analysiswouldbecompletely
detectedin thehemoglobin
invisibleto the mtDNAanalysis.The oldestevent detected in the hemoglobinanalysisis an out-of-Africa
rangeexpansionfoundamong2-stepclades as nested
must
tree,andwhichtherefore
withintheentirehaplotype
haveoccurredclose to the timedepthof the entiretree.
expansioneventis obviouslytoooldto
Thisout-of-Africa
model.Bebytherecentcandelabra
betheonepostulated
cause it spansthe entiretime depthof the hemoglobin
atallabouttheprehaplotypetree,thereis noinfonnation
exexpansionpopulation.Hence,this old out-of-Africa
eventof emptyarpansioncouldhavebeena colonization
eventinwhich
event,orahybridization
eas,areplacement
new migrantsinterbredwithpreviousEurasianinhabitants.Thereis simplyno wayof knowing.Afterthisexpansion,geneflow clearlyoccurredamongAfricansand
Eurasiansas constrainedby isolationby distanceas
shown by the 1-stepcladesnestedwithinboth 2-step
definingthesemid-level
clades(Figure8). Themutations
et
cladesareexpectedto be> 200,000yearsold(Harding
gene
al. 1997).Giventhatthe mtDNAshowsrecurrent
flow with isolation by distance certainly for times
< 200,000yearsago (Figure6), thetwo datasetsjointly
geneticcontactamong
implyalongtimespanof recurrent
themajorOldWorldhumanpopulations.
An out-of-Asiaexpansionevent is detectedwithin
analysis(Figure8).Oneofthe
clade1-5inthehemoglobin
thisexpansionevent(themucriticalmutationsdeElning
tationonthebranchbetweenC3andC2)hasanestimated
ageof 137,000+ 81,500yearsanda secondcriticalmutation (the one definingC7) of 69,000+ 48,000 years
(Hardinget al. 1997).If this out-of-Asiaexpansionis
olderthan100,000,thenitwouldbeimpossibleforacomof theancestralAsianpopulapletegeneticreplacement
tion to haveoccurredby Africans100,000yearsago. If
thisout-of-Asiaexpansionis youngerthan100,000,then
betweenAsiansandAfritherewas geneticinterchange
cans,andthereforeno "split"betweenAfricansandEurasians100,000yearsago. Anotherrangeexpansionis
foundwithinclade 1-2 (Figure8), andthe geographical
impliesthatthisis anout-ofof thehaplotypes
distribution
Africaexpansion.Becauseclade 1-2 includesthe very
oldesthaplotypes,thismaysimplybe a reflectionof the
rangeexpansiondetectedamongthe2old out-of-Africa
step clades.However,the significanteffect of the old
haplotypeB3 mayinthiscasebedueinpartto thenonsigof haplotypeBl . The
distribution
nificantbutwidespread
mutationdefiningB1 has an estimatedage of about

646
SEPTEMBER1 998
152,000 years, but its confidence intervalspans virtually

theentirepast 300,000 years(Hardinget al. 1997). Hence,
the clade 1-2 inferencemay representa more recent outof-Africa expansion occulTing sometime in the last
300,000 years. Even if clade 1-2 representsa recent outof-Africa expansion event, it certainly is not a replacement event. A true replacementevent at about 100,000
years ago would have obliteratedall evidence for older
gene flow; yet the clade l -2 out-of-Africaevent is nestedwithina patternof signiS1cantgene flow with isolation by
distance(clade 2-1). Thereis no obvious way to reconcile
the hemoglobin datawith a recentout-of-Africareplacementevent.
The second factorto keep in mindwhen comparingthe
mtDNA with the hemoglobinarlalysisis the level of dynamic resolution.The nested clade analysis can only detect population events and recurrent forces that are
marked by mutational changes (Templeton l 998a).
MtDNA is evolving muchmorerapidlythanthe hemoglobin locus, and the attendanthaplotypetrees are far more
resolved for mtDNA thanfor hemoglobin.Consequently,
the hemoglobin analysisis on both an older anda coarser
time scale than the mtDNA. Therefore,the most recent
events and forces detectedin the mtDNA analysis would
be invisible to thehemoglobinanalysis.This explainswhy
the hemoglobin analysisdoes not detect the more recent
range expansions revealed by the mtDNA (Figure 6):
thereare simply no or too few mutationsin the hemoglobin datato marktheserecentexpansionevents. Hence,the
hemoglobin and mtDNA analyses are complementary,
not contradictory.
Finally, genetic interchangebetween AfricansandEurasiansis additionallysuggestedby a nested clade analysis of a Y-DNA haplotypetree (Hammeret al. 1998). Interestingly, a range expansion out of Africa and into
Eurasiais detected in this nested analysis. However, in
light of the mtDNA and hemoglobinresults, this expansion was not a replacementevent, at least for the maternal
demographiccomponent.Following this out of Africaexpansion,thenestedanalysisrevealsa patternof signiE1cant
recurrentgene flow restrictedby isolationby distance,including interchangebetweenAfricanandEurasianpopulations. Moreover,there was a subsequentrange expansion out of Asia andintoAfrica,as was also detectedin the
hemoglobin analysis. The Y-chromosome therefore
shows more evidence of long-range population movements thanthe mtDNA. Onepossible explanationfor this
patternis that males dispersedmore thanfemales during
long-range population movements. However, both
mtDNA andY-DNA showrecurrentgene flow with isolation by distance interconnectingAfrican and Eurasian
populations,indicatingthatbothmales and females have
dispersed during short-range migrations. Regardless,
thereis clearly genetic interchangebetween Africansand
Eurasiansdue to a mixtureof gene flow mediatedby isola-
tion by distance and populationmovements. No genetic

split between Africans and Eurasiansis found in the YDNA, as was also true for the mtDNA and hemoglobin
betaregion.
Combined,the mtDNA, Y-DNA, andhemoglobindata
sets reveal that human evolution from about a million
years ago to the last tens of thousandsof years has been
dominated by two evolutionary forces: (1) population
movements and associated range expansions (perhaps
with some local replacements, but definitely with no
globalreplacementwithinthe last 100,000 years),and(2)
gene flow restrictedby isolationby distance.Theonly evidencefor anysplitorfragmentationeventin humanevolutionaryhistorywithin this time frameis the one detected
with mtDNA (Figure6) involving the colonizationof the
Americas(Templeton1998a). However,thiscolonization
was due to eithermultiplecolonizationevents or involved
movements by large numbers of peoples (Templeton
1998a), resultingin extensive sharingof geneticpolymorphisms of New World with Old World humanpopulations.Moreover,thegenetic isolationbetweentheOldand
New Worlds was brief and no longer exists. Otherthan
this temporaryfragmentationevent, the major human
populationshave been interconnectedby gene flow (recurrentat least on a time scale of the orderof tens of thousands of years) duringthe last one to two hundredthousand years. Gene flow may have been more sporadic
earlier, but multiple genetic interchangescertainly occurredamong Old World populations > 200,000 years
ago. Hence, the haplotypeanalyses of geographicalassociationsstronglyrejectthe existence of evolutionarysublineages of humans, reject the separationof Eurasians
from Africans 100,000 years ago, and reject the idea of
"pureraces"in thepast.Thus,human"races"haveno biological validity underthe evolutionarylineage definition
of subspecies.
Conclusions
The geneticdataareconsistentlyandstronglyinformative abouthumanraces.Humansshow only modestlevels
of differentiationamong populationswhen comparedto
otherlarge-bodiedmammals,andthis level of differentiation is well below the usualthresholdusedto identifysubspecies (races)in nonhumanspecies. Hence,humanraces
do not exist underthe traditionalconcept of a subspecies
as being a geographically circumscribed population
showing sharp genetic differentiation.A more modern
definitionof race is thatof a distinctevolutionarylineage
withina species. The genetic evidence stronglyrejectsthe
existence of distinct evolutionary lineages within humans.The widespreadrepresentationof human"races"as
brancheson an intraspecificpopulationtreeis genetically
indefensibleand biologically misleading,even when the
ancestralnode is presentedas being at 100,000yearsago.

TEMPLETON /
Attemptsto salvage the idea of human"races"as evolutionary lineages by invoking greaterracial purityin the
past followed by admixtureevents are unsuccessful and
falsified by multilocuscomparisonsof geographicalconcordanceandby haplotypeanalyses.Instead,all of the genetic evidence shows thatthereneverwas a splitor separation of the "races"or between Africans and Eurasians.
Recent humanevolution has beerlcharacterizedby both
populationrangeexpansions(withperhapssome local replacements but no global replacementwithin the last
100,000 years) and reculTentgenetic interchange.The
100,000 years ago "divergencetime"betweenEurasians
and Africansthatis commonlyfoundin the recentliteratureis really only an "effectivedivergencetime"in sensu
Nei and Roychoudhury(1974, 1982). Since no split occurredbetween AfricansandEurasians,it is meaningless
to assigna dateto an"event"thatneverhappened.Instead,
the effective divergencetime measuresthe amountof restrictedgene flow amongthepopulations(Slatkin1991).
Because of the extensive evidence for genetic interchange through population movements and recurrent
gene flow going back at least hundredsof thousandsof
years ago, there is only one evolutionarylineage of humanity and there are no subspeciesor races undereither
the traditionalor phylogeneticdeElnitions.Humanevolution andpopulationstructurehavebeenandarecharacterized by many locally differentiatedpopulationscoexisting atanygiven time,butwithsufElcientgeneticcontactto
make all of humanitya single lineage sharinga common,
long-terrnevolutionaryfate.
Note
Acknowledgments.
I wouldlike to thankDr.RobertSussman,
Dr. ErikTrinkaus,andthreeanonymousreviewersfor theirexcellent suggestionsfor improvinganearlierdraftof this paper.
References Cited
Aiello, L. C.
1997 Review of "RaceandHumanEvolution:A FatalAttraction."Nature386:350.
Amato,G., andJ. Gatesy
1994 PCR Assays of VariableNucleotideSites for Identification of ConservationUnits. In MolecularEcology and
Evolution:Approachesand Applications.B. Schierwater,
B. Streit,G. P. Wagner,andR. DeSalle, eds. Pp. 215-226.
Basel: BirkhauserVerlag.
Arctander,P., P. W. Kat,B. T. Simonsen,andH. R. Siegismund
1996 Population Genetics of Kenyan Impalasonsequences for Conservation. In Molecular Genetic Approachesin Conservation.T. B. Smith and R. K. Wayne,
eds. Pp. 399-412. Oxford:OxfordUniversityPress.
647
Armour,J. A. L., T. Anttinen,C. A. May,E. E. Vega, A.

Sajantila,J. R. Kidd,K. K. Kidd,J. Bertranpetit,S. Paabo,and
A. J. Jeffreys
1996 Minisatellite Diversity Supports a Recent African
OriginforModernHumans.NatureGenetics 13:154-160.
Arnason,U., A. Gullberg,A. Janke,andX. Xu
1996 Pattern and Timing of Evolutionary Divergences
among Hominoids Based on Analyses of Complete
mtDNAs.Journalof MolecularEvolution43:65>661.
Avise, J. C.
1994 MolecularMarkers,NaturalHistory and Evolution.
New York:ChapmanandHall.
Barbujani,G., A. Magagni,E. Minch,andL. L. Cavalli-Sforza
1997 An Apportionmentof Human DNA Diversity. Proceedings of the National Academy of Sciences USA
94:451S4519.
Barton,N. H., andS. Rouhani
1993 Adaptationandthe "ShiftingBalance."GeneticalResearch61:57-74.
Batzer,M. A., M. Stoneking,M. Alegria-Hartman,
H. Bazan,
D. H. Kass,T. H. Shaikh,G. E. Novick, P. A. Ioannou,W. D.
Scheer,R. J. Herrera,andP. L. Deininger
1994 AfricanOrigin of Human-SpecificPolymorphicAlu
Insertions.Proceedingsof the National Academy of Sciences USA 91: 12288-12292.
Bowcock, A. M., J. R. Kidd,J.L. Mountain,J. M. Hebert,L.
Carotenuto,K. K. Kidd,andL. L. Cavalli-Sforza
1991 Drift,Admixture,andSelectionin HumanEvolution:
A Study with DNA Polymorphisms.Proceedingsof the
NationalAcademyof Sciences USA 88:839-843.
Bowcock, A. M., A. Ruiz-Linares,J.Tomfohrde,E. Minch,J.
R. Kidd,andL. L. Cavalli-Sforza
1994 High Resolution of HumanEvolutionaryTrees with
PolymorphicMicrosatellites.Nature368:455-457.
Boyce, W. M., P. W. Hedrick,N. E. Mugglicockett,S.
Kalinowski,M. C. T. Penedo,andR. R. Ramey
1997 Genetic Variationof MajorHistocompatibilityComplex andMicrcosatelliteLoci- A Comparisonin Bighorn
Sheep.Genetics 145:421-433.
Brownlow,C. A.
1996 Molecular Taxonomy and the Conservationof the
Red Wolf and Other EndangeredCarnivores.Conservation Biology 10:39(}396.
Calafell,F., P. Underhill,A. Tolun,D. Angelicheva,andL.
Kalaydjieva
1996 FromAsia to Europe:MitochondrialDNA Sequence
Variabilityin BulgariansandTurks.Annalsof HumanGenetics 60:35-49.
Cann,R. L., M. Stoneking,andA. C. Wilson
1987 MitochondrialDNA and Human Evolution. Nature
325:31-36.
Cavalli-Sforza,L., P. Menozzi, andA. Piazza
1996 TheHistoryandGeographyof HumanGenes.Princeton, NJ:PrincetonUniversityPress.
Cavalli-Sforza,L. L.
1997 Genes, Peoples, and Languages.Proceedingsof the
NationalAcademyof Sciences USA 94:7719-7724.

648
* SEPTEMBER1998
Comas,D., F. Calafell,E. Mateu,A. Perez-Lezaun,E. Bosch,

andJ. Bertranpetit
1997 MitochondrialDNA Variationand the Origin of the
Europeans.HumanGenetics99:443-449.
Cooper,G., W. Amos, D. Hoffman,andD. C. Rubinsztein
1996 NetworkAnalysis of HumanY MicrosatelliteHaplotypes. HumanMolecularGenetics5:1759-1766.
Crow,J. F., andM. Kimura
1970 An Introductionto PopulationGenetic Theory.New
York:HarperandRow.
Deka, R., L. Jin,M. D. Shriver,L. M. Yu, S. DeCroo,J.
Hundrieser,C. H. Bunker,R. E. Ferrell,andR. Chakraborty
1995 PopulationGenetics of Dinucleotide (dC-dA)n.(dGdT)nPolymorphisms in World Populations. American
Journalof HumanGenetics56:461-474.
DeSalle, R., A. Templeton,I. Mori,S. Pletscher,andJ. S.
Johnston
1987 TemporalandSpatialHeterogeneityof mtDNA Polymorphismsin NaturalPopulationsof Drosophila mercatorum.Genetics 116:215-223.
Ellsworth,D. L., R. L. Honeycutt,N. J. Silvy, J. W. Bickham,
andW. D. Klimstra
1994 HistoricalBiogeographyand ContemporaryPatterns
of MitochondrialDNA Variation in White-Tailed Deer
from the SoutheasternUnited States. Evolution 48:122136.
Endler,J. A.
1977 Geographic Variation, Speciation, and Clines.
Princeton,NJ:PrincetonUniversityPress.
Francalacci,P., J. Bertranpetit,F. Calafell,andP. A. Underhill
1996 Sequence Diversity of the Control Region of Mitochondrial DNA in Tuscany and Its Implications for the
Peoplingof Europe.AmericanJournalof PhysicalAnthropology 100:443-460.
Futuyma,D. J.
1986 EvolutionaryBiology. Sunderland,MA: SinauerAssociates, Inc.
Georgiadis,N., L. Bischof, A. Templeton,J. Patton,W.
Karesh,andD. Western
1994 Structureand History of African Elephant Populations: I. Easternand SouthernAfrica. Journalof Heredity
85:10W104.
Hammer,M. F., A. B. Spurdle,T. Karafet,M. R. Bonner,E. T.
Wood, A. Novelletto,P. Malaspina,R. J. Mitchell,S. Horai,T.
Jenkins,andS. L. Zegura
1997 The GeographicOistributionof tIuman Y Chromosome Variation.Genetics145:787-805.
Hammer,M. F., T. Karafet,A. Rasanayagam,E. T. Wood, T.
K. Altheide,T. Jenkins,R. C. Griffis, A. R. Templeton,and
S. L. Zegura
In press Out of Africa and Back Again: Nested Cladistic
Analysis of HumanY ChromosomeVariation.Molecular
Biology andEvolution.
Harding,R. M., S. M. Fullerton,R. C. Griffiths,J. Bond, M. J.
Cox, J. A. Schneider,D. S. Moulin,andJ. B. Clegg
1997 Archaic Africanand Asian Lineages in the Genetic
Ancestryof ModernHumans.AmericanJournalof Human
Genetics60:772-789.
Harpending,H. C., J. H. Relethford,andS. T. Sher}y

1996 Meffiodsand Models for UnderstandingHumanDiversity.In MolecularBiology andHumanDiversity.A. J.
Boyce and C. G. N. Mascie-Taylor,eds. Pp. 283-299.
Cambridge:CambridgeUniversityPress.
Hey, J.
1997 Mitochondrialand Nuclear Genes PresentConflicting Portraitsof Human Origins. MolecularBiology and
Evolution 14:16S172.
Hollocher,H., andA. R. Templeton
1994 The Molecular throughEcological Genetics of AbVI. The
nonnal Abdomen in Drosophilamercatorum
Nonneutrality of the Y-chromosome rDNA Polymorphism.Genetics 136:1373-1384.
Hollocher,H., A. R. Templeton,R. DeSalle, andJ. S. Johnston
1992 The Molecular throughEcological Genetics of AbnormalAbdomen.IV. Componentsof Genetic-Variation
Gein a Natural-Populationof Drosophilamercatorum.
netics 130:355-366.
Howell, N., I. Kubacka,andD. A. Mackey
1996 How Rapidly Does the Human Mitochondrial
Genome Evolve? American Journalof Human Genetics
59:501-509.
Hudson,R. R., D. D. Boos, andN. L. Kaplan

1992 A StatisticalTest for Detecting GeographicalSubdivision. MolecularBiology andEvolution9:138-15 1.
Jorde,L. B., M. J. Bamshad,W. S. Watkins,R. Zenger,A. E.
Fraley,P. A. Krakowiak,K. D. Carpenter,H. Soodyall,T.
Jenkins,andA. R. Rogers
1995 Origins and Affinities of ModernHumans:A Comparison of Mitochondrial and Nuclear Genetic Data.
AmericanJournalof HumanGenetics57:523-538 .
Jorde,L. B., A. R. Rogers,M. Bamshad,W. S. Watkins,P.
Krakowiak,S. Sung, J. Kere,andH. C. Halpending
1997 MicrosatelliteDiversity and the DemographicHistory of Modern Humans. Proceedings of the National
Academyof Sciences USA 94:3100-3103 .
King,R. B., andR. Lawson
1995 Color-Pattern Variation in Lake Erie Water
Snakes the Role of Gene Flow. Evolution49:885-896.
1997 Microevolutionin Island WaterSnakes. BioScience
47:279-286.
Krings,M., A. Stone, R. W. Schmitz,H. Krainitzki,M.
Stoneking,andS. Paabo
1997 NeandertalDNA sequencesandtheOriginof Modern
Humans.Cell 90:19-30.
Lasker,G. W., andD. E. Crews
1996 Behavioral Influences on the Evolution of Human
GeneticDiversity.MolecularPhylogeneticsandEvolution
5:232-240.
Lawson,R., andR. B. King
1996 Gene Flow and Melanismin LakeErie GarterSnake
Populations. Biological Journalof the Linnean Society
59: 1-1 9.
Legge, J. T., R. Roush,R. Desalle, A. P. Vogler,andB. May

1996 Genetic Criteriafor EstablishingEvolutionarilySignificantUnits in CryansBuckmoth.ConservationBiology
10:85-98.

GENETICS
OFRACE
Lehman,N., andR. K. Wayne

1991 Analysis of Coyote Mitochondrial-DNAGenotype
Frequencies:Estimationof the Effective Number of Alleles.Genetics 128:405-416.
Lynch,M., andT. J. Crease
1990 The Analysis of PopulationSurveyData on DNA Sequence Variation. Molecular Biology and Evolution
7:377-394.
Maddison,D. R.
1991 AfricanOriginof HumanMitochondrialDNA Reexamined.SystematicZoology 40:355-363.
Manderscheid,E. J., andA. R. Rogers
1996 GeneticAdmixturein theLatePleistocene.American
Journalof PhysicalAnthropology100:1-5.
Mayr,E.
1970 Populations, Species, and Evolution. Cambridge,
MA: BelknapPressof HarvardUniversityPress.
Mitchell,R. J.
1996 Y-Chromosome-Specific
RestrictionFragmentLength
Polymorphisms (RFL,Ps)-Relevance to Human Evolution and Human Variation.AmericanJournalof Human
Biology 8:573-586.
Miththapala,S., J. Seidensticker,andS. J. O'Brien
1996 Phylogeographic Subspecies Recognition in Leopards (Panthera pardus9Molecular Genetic Variation.
ConservationBiology 10:1115-1 132.
MountainsJ. L., andL. L. Cavalli-Sforza
1994 Inference of Human Evolution through Cladistic
Analysis of Nuclear DNA Restriction Polymorphisms.
Proceedings of the National Academy of Science USA
91:6515-6519.
Nachman,M. W., W. M. Brown,M. Stoneking,andC. F.
Aquadro
1996 Nonneutral MitochondrialDNA Variation in HumansandChimpanzees.Genetics 142:953-963.
Nei, M., andA. K. Roychoudhury
1974 Genic Variationwithinandbetweenthe ThreeMajor
Races of Man: Caucasoids, Negroids, and Mongoloids.
AmericanJournalof HumanGenetics26:421-443.
1982 GeneticRelationshipandEvolutionof HumanRaces.
EvolutionaryBiology 14:1-59.
1993 EvolutionaryRelationshipsof HumanPopulationson
a Global Scale. Molecular Bio]ogy and Evolution
10:927-943.
Nei, M., andN. Takezaki
1996 lshe Root of the PhylogeneticTreeof HumanPopulations. MolecularBiology andEvolution13:17>177.
Parsons,T. J., D. S. Muniec,K. Sullivan,N. Woodyatt,R.
Allistongreiner,M. R. Wilson,D. L. Berry,K. A. Holland,V.
W. Weedn,P. Gill, andM. M. Holland
1997 A High ObservedSubstitutionRatein theHumanMitochondrial DNA Control Region. Nature Genetics
15:363-368.
Pennock,D. S., andW. W. Dimmick
1997 Critiqueof the EvolutionarilySigniElcantUnit as a
Definition for Distinct Population Segments under the
U.S. Endangered Species Act. Conservation Biology
11:611-619.
649
Perez-Lezaun,A., F. Calafell,E. Mateu,D. Comas,R.

Ruiz-Pacheco,andJ. Bertranpetit
1997 Microsatellite Variation and the Differentiationof
ModemHumans.HumanGenetics99:1-7.
Relethford,J. H., andH. C. Harpending
1994 CraniometricVariation,GeneticTheory,andModern
HumanOrigins.AmericanJournalof Physical Anthropology 95:249-270.
1995 AncientDifferences in PopulationSize CanMimic a
Recent African Origin of ModernHumans. CurrentAnthropology36:667-674.
Rogers,A. R.
1997 PopulationStructureand ModernHumanOrigins.In
Progressin PopulationGenetics andHumanEvolution.P.
Donnelly andS. Tavare,eds. Pp. 55-79. New York:Springer.
Rohlf,F. J.
1993 NTSYS-pc: Numerical Taxonomy and Multivariate
AnalysisSystem,Version 1.80. Setauket,NY: ExeterSoftware.
Santos,E. J. M., J.T. Epplen,andC. Epplen
1997 Extensive Gene Flow in HumanPopulationsas Revealed by Protein and MicrosatelliteDNA Markers.HumanHeredity47:165-172.
Shaffer,H. B., andM. L. McKnight
1996 The Polytypic Species Revisitedenetic
DifferentiationandMolecularPhylogeneticsof the TigerSalamander Ambystomatigrinum(Amphibia,Caudata)Complex.
Evolution50:417-433.
Slatkin,M.
1991 InbreedingCoefficients and CoalescenceTimes. GeneticalResearch58:167-175.
Smith,H. M., D. Chiszar,andR. R. Montanucci
1997 Subspecies and Classification. Herpetological Review 28:13-1 6.
Stoneking,M.
1997 Recent African Origin of Human Mitochondrial
DNA: Review of the Evidence and CurrentStatusof the
Hypothesis.In Progress in PopulationGenetics and Human Evolution.P. Donnelly and S. Tavare,eds. Pp. 1-13.
New York:Springer.
Tavare,S. D. Balding,R. C. Griffiths,andP. Donnelly
1997 InferringCoalescence Times from DNA Sequence
Data.Genetics 145:505-518.
Templeton,A. R.
1980 The Theoryof Speciationvia the FounderPrinciple.
Genetics94:1011-1038.
1992 HumanOriginsand Analysis of MitochondrialDNA
Sequences.Science 255:737.
1993 The "Eve"Hypothesis: A Genetic Critiqueand Reanalysis.AmericanAnthropologist95:51-72.
1994a "Eve":HypothesisCompatibilityversusHypothesis
Testing.AmericanAnthropologist96:141-1 47.
1994b The Role of MolecularGenetics in SpeciationStudies. In MolecularEcology andEvolution:Approachesand
Applications.B. Schierwater,B. Streit,G. P. Wagner,and
R. DeSalle, eds. Pp. 455-477. Basel: Birkhauser-Verlag.
1996 Contingency Tests of Neutrality Using Intra/Interspecific Gene Trees: The Rejection of Neutralityfor the

650
1998
* VOL. 100, NO. 3 * SEPTEMBER
ANTHROPOLOGIST
AMERICAN
Evolution of ie MitochondrialCytochrome Oxidase II

Gene in theHominoidPrimates.Genetics 144:1263-1270.
1997a Out of Africa? What Do Genes Tell Us? Current
Opinionin GeneticsandDevelopment7:841-847.
1997b Testing the Out-of-AfricaReplacementHypothesis
with MitochondrialDNA Data. ln ConceptualIssues in
ModernHumanOriginsResearch.G. A. Clarkand C. M.
Willermet, eds. Pp. 329-360. New York: Aldine de
Gruyter.
1998a Nested Clade Analyses of PhylogeographicData:
TestingHypothesesaboutGene Flow andPopulationHistory.MolecularEcology 7:381-397.
1998 Species and Speciation: Geography, Population
Structure,Ecology, and Gene Trees. In Endless Forms:
Species andSpeciation.D. J. Howardand S. H. Berlocher,
eds. Pp. 32-43. Oxford:OxfordUniversityPress.
Templeton,A. R., E. Boerwinkle,andC. F. Sing
1987 A CladisticAnalysisof PhenotypicAssociationswith
HaplotypesInferredfrom RestrictionEndonucleaseMapping. I. Basic Theoryandan Analysisof Alcohol Dehydrogenase Activity in Drosophila.Genetics 117:343-351 .
Templeton,A. R., andN. J. Georgiadis
1996 A LandscapeApproach to ConservationGenetics:
ConservingEvolutionaryProcesses in the African Bovidae. In ConservationGenetics: Case Histories from Nature.J. C. Avise andJ. L. Hamrick,eds. Pp. 398-430. New
York:ChapmanandHall.
Templeton,A. R., H. Hollocher,andJ. S. Johnston
1993 The MolecularthroughEcological Genetics of AbnormalAbdomen in Drosophila mercatorum.V. Female
PhenotypicExpressionon NaturalGenetic Backgrounds
andin NaturalEnvironments.Genetics 134:475-485.
Templeton,A. R., H. Hollocher,S. Lawler,andJ. S. Johnston
1989 Natural Selection and Ribosomal DNA in Drosophila. Genome31:296-303.
Templeton,A. R., E. Routman,andC. Phillips
1995 Separating Population Structure from Population
History:A CladisticAnalysis of the GeographicalDistributionof MitochondrialDNA Haplotypesin the TigerSalAmbystomatigrinum.Genetics 140:767-782.
arnander,
Templeton,A. R., andC. F. Sing

1993 A CladisticAnalysis of PhenotypicAssociationswith
HaplotypesInferredfrom RestrictionEndonucleaseMapping. IV. Nested Analyses with CladogramUncertainty
andRecombination.Genetics 134:659-669.
TexTell,J.E., andP. J. Stewart
1996 The Paradox of Human PopulationGenetics at the
End of the Twentieth Century.Reviews in Anthropology
25:13-33.
Vigilant,L., M. Stoneking,H. Harpending,K. Hawkes,andA.
C. Wilson
1991 AfricanPopulationsandthe Evolutionof lIumanMitochondrialDNA. Science 253:1503-1507.
Vogler,A. P.
1994 Extinction and the Formationof PhylogeneticLineages: Diagnosing Units of ConservationManagementin
the Tiger Beetle Cicindeladorsalas.In PCR Assays of
VariableNucleotide Sites for Identificationof Conservation Units. B. Schierwater,B. Streit,G. P. Wagner,andR.
DeSalle, eds. Pp. 261-273. Basel:BirkhauserVerlag.
WatsonsE., K. Bauer,R. Aman,G. Weiss, A. vanHaeseler,
andS. Paabo
1996 mtDNA Sequence Diversity in Africa. American
Journalof HumanGenetics59:437-444.
Wayne,R. K., N. Lehman,M. W. Allard,andR. L. Honeycutt
1992 MitochondrialDNA Variabilityof the Gray Wolf:
Genetic Consequencesof PopulationDecline andHabitat
Fragmentation.ConservationBiology 6:559-569.
Wilson,E. O., andW. L. Brown
1953 The Subspecies ConceptandIts TaxonomicApplications. SystematicZoology 2:97-1 11.
Wolpoff,M., andR. Caspari
1997 Race and HumanEvolution. New York: Simon and
Schuster.
Wright,S.
1931 Evolution in Mendelian Populations.Genetics 16:
97-1 59.
1943 Isolationby Distance.Genetics28:11F138.
1969 Evolution and the Genetics of Populations.3 vols.
Volume 2. The Theory of Gene Frequencies. Chicago:
Universityof ChicagoPress.


Templeton Human Races-2

Загружено:

Сведения о документе

Исходное описание:

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Templeton Human Races-2

Загружено:

Авторское право:

Доступные форматы

Human Races: A Genetic and Evolutionary Perspective

Author(s): Alan R. Templeton

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

HumanRaces:A Geneticand EvolutionaryPerspective

word raceis rarelyused in the modern,nonhumanevolutionaryliteraturebecauseits meaningis

AmericanAnthropologist100(3):632-650. Copyright(3 1999, AmericanAnthropologicalAssociation

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

is no need to prioritizethe informativetraitsin defining

abilities.Figure 1 shows thevalues of Fstsandrelatedstatistics for severallarge-bodiedmammals.As can be seen,

Are Human "Races" Geographically

Most real populationsdo not fit an island model (which

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

B. Recent Origin Candelabra

mainderof this paper will focus upon this more modern

Are Human "Races" Distinct Evolutionary

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

AMERICANANTHROPOLOGIST * VOL. 100, NO. 3

Figure 3. The trellis model of recent human evolution. Under this

candelabra(Figure2) and trellis (Figure3) models. Both

nized Eurasiaabout 100,000 yearsago, drivingthe Homo

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

Genetic Diversity Levels within and among

whereNeis an effective size of the populationand,uis the

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

AMERICANANTHROPOLOGIST * VOL. 100, NO. 3

growth which causes an increase not a decrease in

C. A. R. Zaire Europeans Chinese Melanesians

Genetic Distances and Evolutionary "Trees"

Figure 4. Genetic distances and recent human evolution. Part A

tween Africanandnon-Africanpopulationshavethe same

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

of what evolutionaryfactorsgeneratedthe distances.It is

Geographic Distance in Miles

hensive humandataset andconcludedthat"theisolationby-distancemodels hold for long distancesas well as for

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

For example, consider Melanesiansand Africans. As

bornnearbybut often outsideone's own natalgroup

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

possible to estimatea networkthatshows how mutational

mitochondrialDNA can spread only when populations

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

AMERICANANTHROPOLOGIST * VOL. 100, NO. 3

effective size, andsuchmutationsareknownto exist (Hey

type networkis then prunedoff and the process repeated

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

the clades that markthose geographicalregions and that

significant results summarizedin Figure 6 (Templeton

Restricted Gene Flow After

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

AMERICANANTHROPOLOGIST * VOL. 100, NO. 3

sign, the geographical data are quantifiedin two main

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

1-2-34 No: IBD

1-2-34 No: IBD

Figure 8. Resultsof the nestedgeographicanalysisof thehumanbeta

referredto as the TRP key). Templeton(1998a) gives an

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

AMERICANANTHROPOLOGIST * VOL. 100, NO. 3

152,000 years, but its confidence intervalspans virtually

tion by distance and populationmovements. No genetic

This content downloaded from 132.236.27.111 on Sat, 28 Dec 2013 09:29:09 AM

Armour,J. A. L., T. Anttinen,C. A. May,E. E. Vega, A.