The Peopling of the Americas

Author(s): Ann Gibbons

Source: Science, New Series, Vol. 274, No. 5284 (Oct. 4, 1996), pp. 31-33
Published by: American Association for the Advancement of Science
Stable URL: http://www.jstor.org/stable/2891831
Accessed: 17-09-2016 00:03 UTC

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.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

American Association for the Advancement of Science is collaborating with JSTOR to digitize,
preserve and extend access to Science

This content downloaded from 157.182.150.22 on Sat, 17 Sep 2016 00:03:14 UTC
All use subject to http://about.jstor.org/terms
- RESEARCH NEWS

The Peopling of the Americas

New genetic data suggest that the earliest Americans came from Asia in one or two waves-not
more-challenging an earlier synthesis of linguistic, dental, and genetic evidence

Six years ago, D. Andrew Merriwether was a

European team whose review is published in and Aleutian Islanders), and Na-Dene (spo-
master's student in the lab of geneticist the October issue of the American Journal ken by people of the Northwest coast of
Douglas Wallace at Emory University in At- of Human Genetics. Additional DNA samples Canada and the United States). There were
lanta, learning to use genes to trace the an- and better resolution show that native also three types of molar shapes and three
cestry of native American peoples. When he Americans as diverse as the Eskimos of genetically distinct populations. "Every time
left to pursue doctoral studies at the Univer- Alaska and the Kraho and Yanomami of you come around, it's three," says Arizona
sity of Pittsburgh, Merriwether continued Brazil share more gene types than previ- State University bioarchaeologist Christy
that research, expecting to bolster the con- ously thought, which suggests that they are Turner, co-author of the hypothesis.
clusion coming from the Wallace lab: that descended from the same founding popula- The dates also meshed with existing ar-
genetically distinct groups of prehistoric tions in Asia-and that their ancestors en- chaeological data. Using the degree of differ-
people migrated to the Americas in three tered North America in only one or two ence among languages, Greenberg calculated
separate waves. But this year, Merriwether migratory waves, says Oxford University that the first language arrived in Alaska about
found himself publicly contradicting his evolutionary geneticist Ryk Ward. Scien- 12,000 years before the present. That fits in
mentor, in a series of papers suggesting thattists are already searching for those ances- with the first widely accepted archaeological
there was only a single migration. Although tors' closest kin in Siberia and Mongolia. evidence of culture in the Americas, the
0
11,500-year-old sites of the Clovis people.
At first, the genetic evidence seemed to
tie in, too. Indeed, in the early 1990s, some
6
of the best supporting evidence for the
Greenberg hypothesis came from genes
(Science, 15 January 1993, p. 312). In gen-
eral, the more similar genes are among two
* Eskimo-Aleut
speakers populations, the more closely the popula-
* Na-Dene speakers \*s tions are related. To trace these similarities
* Amerind speakers
Alternate routes. The in native American DNA, Wallace, work-
Greenberg hypothesis ing with geneticist Antonio Torroni, now
(above) shows three at the University of Rome, and graduate
waves of migration from
Not surprisingly, not everyone supports student Theodore Schurr, assembled hun-
Asia to the Americas,
while the emerging new
the new interpretations. Greenberg, for ex- dreds of blood samples from 24 tribes from
view (right) sees only ample, says that given the flip-flopping con- Alaska to Argentina. They analyzed the
one or two migrations. clusions from the DNA data, he's simply ig- DNA coiled in the mitochondria, the en-
noring it until geneticists reach consensus. ergy factories of the cell. Most anthropo-
they remain personally friendly, mentor and Others caution against putting too much logical studies use this mtDNA because it
student clearly are divided on this issue. Says weight on any one type of genetic data, and mutates faster than nuclear DNA, allowing
Merriwether, now at the University of Wallace still concludes that native Ameri- researchers to distinguish populations that
Michigan: "I feel badly about it because cans arrived in three migrations. He's philo- recently separated. The mtDNA is also in-
Wallace is the one who inspired me to go sophical about the new work, saying that herited only from mothers and so avoids the
into this field. It's awkward." "testing new hypotheses is what research is gene shuffling that can obscure the evolu-
Chalk up one more disagreement to one of all about." But if the new studies are right, tionary trail of most nuclear genes.
the most contentious issues in human prehis-the notion of three or more separate migra- Wallace's group used particular enzymes to
tory: the question of who settled the Ameri-tions is unlikely, and the whole idea of mar- cut the DNA into standard pieces, then
cas. A decade ago, intellectual battles raged rying linguistic and genetic data comes into looked for variations in the length of those
over a bold synthesis of linguistic, genetic, andquestion (see box). "This tends to confirm segments-called restriction fragment length
dental data named after co-creator Joseph our conclusion that there isn't a relationship polymorphisms (RFLPs)-to indicate the
Greenberg, a Stanford University linguist. between genetic signatures of migrations and presence of mutations. They got intriguing
The Greenberg theory suggested that the first language," says Ward. results, finding that native Americans carry
Americans arrived from Asia in at least three only four variants of mtDNA, called haplo-
separate waves, each wave giving rise to one of Evolutionary triangle groups A, B, C, and D, with each group char-
three linguistic groups. Linguists opposed put- The Greenberg hypothesis, although contro- acterized by a different set of mutations. These
ting the diverse languages of most native versial, had great appeal because it synthe- variants were found in some East Asians and
Americans into one "Amerind" group, but the sized so many independent lines of evidence. Siberians (but not in Europeans or Africans),
theory fit dental and genetic evidence from And no matter what one was comparing- which indicated that the mutations originally
several labs, including Wallace's. languages, teeth, or genes-the magic num- came from Asia. Not every indigenous group
But now the pillar of support from genet- ber was three. There were three linguistic seemed to carry all four, however. Wallace's
ics is showing cracks, thanks to new data groupings: Amerind (spoken by American team found, for example, that although most
from Merriwether and others, including a Indians), Eskimo-Aleut (spoken by Eskimos Amerind speakers carried all four haplo-

SCIENCE * VOL. 274 * 4 OCTOBER 1996 31

This content downloaded from 157.182.150.22 on Sat, 17 Sep 2016 00:03:14 UTC
All use subject to http://about.jstor.org/terms
groups, Na-Dene speakers Diverse descen- carried all four haplogroups.
carried just one (haplogroup < dants. Greenberg pro- Once in America, this first wave of set-
posed that the ances-
A), and the Eskimo-Aleut tlers spread out. Some pushed south, but oth-
tors of these peoples
speakers carried two (haplo- came to the Americas
ers stayed in the northwest, where their num-
groups A and D). So the in three separate bers were drastically reduced-perhaps by
team concluded that Amer- < waves: Eskimos (Es- bitter cold during the last glacial period that
ind speakers descended from kimo-Aleut speakers, ended about 1 1,500 years ago. As a result, the
women who carried all four left); the Maya (Amer- northern populations, the ancestors of the
ind speakers, bottom
types, while the other two Na-Dene and Eskimo-Aleuts, lost their
left); and the Tlingit
groups descended from wo- original genetic diversity. Their numbers
(Na-Dene speakers,
men who carried just one or bottom right). eventually bounced back, but with fewer
two. And this suggested that copies of haplogroups B, C,
z cn
they came to the New World and D than carried by their
in three distinct waves from 0
CID
0
-0
southern relatives.
Asia, just as Greenberg had I
o~~~~~~~~~ The latest word on the
proposed. I
settling of the Americas
comes from Europe-and it
I
New-Wave models 0
iL
-~~~~~~~~~~~~~~~~~~~~~~~~~~~C too challenges the three-

__ ~~~~~~~~
But just as it seemed a con- Ye
migration theory. An inter-
sensus was emerging, genet- xI _ - f f~~~~~~~~~~~~~~~~i disciplinary group tackled
icists cast their net wider, the problem by re-analyzing
testing the DNA of more na- recent data on mtDNA se-
tive Americans and Asians quences, pooling studies of
and, in some cases, looking DNA from a total of 5 74 na-
directly at DNA sequences. The presence of all four tive Americans and Siberians, says molecu-
Merriwether, for example, markers in each of the three lar biologist Peter Forster, a graduate student
was analyzing DNA samples linguistic groups makes it working with Hans-Jurgen Bandelt of Ham-
of 1300 American Indians and other native unlikely that the groups' ancestors came in burg University, Rosalind Harding of the In-
groups, and he kept finding people, such as thedifferent migrations thousands of years apart, stitute of Molecular Medicine at Oxford, and
Yanomami of Brazil, whose genes didn't fit in says Merriwether. "Think of the source popu- Torroni in Rome.
the four lineages identified by Wallace's team.lation as a bowl of colored marbles," says Instead of looking for markers that may
Not only were there more than four genetic Smithsonian Institution molecular anthro- accompany mutations, this team looked di-
variants, but the four original types showed uppologist Connie Kolman. "You won't pick rectly at the DNA sequences-a slower but
in all three major language groups. Merri- out the [same combination ofl rare types more certain way to find variations. They
wether's work was confirmed by a group of three of four times if you reach in randomly." entered mtDNA sequences into a computer
South American researchers, led by Nestor A. So Merriwether and University of Pittsburgh and searched for matches between Ameri-
Bianchi of the Multidisciplinary Institute of geneticist Robert Ferrell joined forces with can tribes, Asians, and Siberians. Ironi-
Cellular Biology in Argentina, who analyzed Francisco Rothhammer of the University of cally, they found all four of the original vari-
mtDNA from 25 populations and observed Chile to propose just a single migration, in ants in almost all the Amerinds-support-
the same results. which the first women to set foot in America ing the part of the Greenberg hypothesis

Can This Marriage Be Saved?

In 1986, a linguist, an archaeologist, and a geneticist gotthe family tree of Amerind languages back to one 12,000-year-old
together
and proposed a match between languages and genes in native ancestor, because written records go back only 5000 years.
American groups. The brainchild of their interdisciplinary union Even if the language patterns are real, they may not imply
was the Greenberg hypothesis, named for Stanford University anything about the genetic relationships of the people who speak
linguist Joseph Greenberg. They proposed that three separate them. "Most linguists don't think that there's a correlation between
migrations from Asia gave rise to three groups distinct in both genes and language," says Thomason. Counters Smithsonian Insti-
genes and language: the American Indians of North and South tution anthropologist Connie Kolman: "I don't think languages can
America, who speak Amerind languages; the Eskimos and Aleuts, tell us anything about the initial colonization of the Americas."
who speak Eskimo-Aleut; and a third group including other Indeed, geneticists are divided on whether people in well-
Northwestern tribes such as the Haida and Tlingit, who speak established language groups show matching genetic patterns in the
languages in the Na-Dene family. Americas. Stanford University geneticist L. Luca Cavalli-Sforza, for
But as one member of the union runs into trouble-the genetic example, is convinced that such correlations exist, based on a wide
evidence (see main text)-many linguists say that it was a shot- range of DNA, chromosomal, and blood-group data. But others
gun marriage from the start. Historical linguists say they never aren't so sure. Oxford University evolutionary geneticist Ryk Ward is
agreed that native American languages fall into the three broad testing the notion by studying the DNA of Na-Dene speakers and
groups Greenberg identified. Over the years, they have poked Eskimo-Aleuts-linguistic groups thought to be quite distinct. But
holes in Greenberg's data showing similarities among the differ- his data indicate that genetically, these groups are "remarkably simi-
ent Amerind languages, says linguist Sarah G. Thomason of the lar to each other." If that finding holds up, says Ward, Greenberg's
University of Pittsburgh. Linguists also say it's impossible to trace interdisciplinary union will be in trouble. -A.G.

32 SCIENCE * VOL. 274 * 4 OCTOBER 1996

This content downloaded from 157.182.150.22 on Sat, 17 Sep 2016 00:03:14 UTC
All use subject to http://about.jstor.org/terms
 RESEARCH NEWS

that linguists dislike most, namely that the group and from several new South Ameri- to a different conclusion. I've just tended to
ancestors of all the Amerinds came in one can archaeological sites (Science, 19 April, set aside the mtDNA evidence. I'll wait until
wave. Overall, their more powerful method p. 346). The re-expansion, they say, hap- they get their act together."
detected nine founding mtDNA sequences pened about 11,300 years ago-the time of Even some geneticists are reluctant to
in native American peoples, and some of the Clovis people. claim they have solved the problem of the
these sequences were only in Na-Dene peopling of the Americas. "I am worried
speakers, Eskimos, and coastal Siberians, In sug-
search of a homeland that too much weight is being given to mi-
gesting that those groups emerged from a If all of today's native Americans do go back tochondrial DNA," says Stanford Univer-
common ancestral population, not from to a single population in Asia, which one? sity geneticist Luigi Luca Cavalli-Sforza. He
separate groups, as Wallace had proposed. The multiple-migration advocates put their notes that mtDNA reflects only the move-
The team put this data together and pro- founders in Siberia, as does the European ments of women. Because women in some
posed that the ancestors of the Amerinds team, because Siberians share some founding hunter-gatherer societies join their hus-
came in the first wave from northeastern Si-variants with Na-Dene speakers and Eski- bands' families and move more than men,
beria and carried all the variants, some of mos-and live close to the land bridge to the their mtDNA may not reveal the migra-
which were lost in northern Asians and Americas. Merriwether and Kolman are skep- tions of the whole population. So Cavalli-
Americans, perhaps due to climate. Later, tical, however, because all modem Siberians Sforza and colleague Peter Underhill, as
the survivors rebounded, probably in Beringia, tested so far lack haplogroup B. In separate well as other teams, are studying markers on
and gave rise to the Na-Dene and Eskimos. papers, Merriwether, and Kolman and her the paternally inherited Y chromosome. So
This scenario allows for either one or two Smithsonian colleague Eldredge Bermingham far, their results don't rule out additional
migrations into North America, depending have proposed that the founders may have migrations.
on whether the homeland of the surviving been Mongolians, because they carry all four Indeed, researchers warn that more data
northerners was in North America or Sibe- haplogroups. from several genetic lineages will be needed
ria. Forster says: "We call it a re-expansion. Even on the number of migrations, there to provide a picture of the peopling of the
It's a matter of taste whether you call it a is no consensus. Satoshi Horai of the Na- Americas. However, that time may not be so
separate migration." tional Institute of Genetics in Mishima, Ja- far off. One of the other authors of the
The European group also explored an pan, for example, notes that his analyses of Greenberg hypothesis, University of Arizona
even more controversial issue, the timing the of genetic distance among native American geneticist Stephen Zegura, is taking the new
these migrations, by using the amount of peoples suggests that there are four groups studies seriously and has taken a sabbatical
genetic difference among populations as a that have been isolated for a relatively long this year to try to sort out the findings: "I'm
molecular clock. The Amerind speakers show period of time. He concludes that there were trying to decide if after 10 years, is the time
the most diversity, so the team concluded four separate migrations. right to do a new synthesis?" The answer
that they arrived in the first wave, 20,000 to All this disagreement prompts Greenberg from a new wave of young geneticists, at
25,000 years ago. That predates the Clovis to simply ignore the new mtDNA data. He least, is a resounding yes.
culture but matches dates from Wallace's says: "Every time, it [mtDNA] seems to come -Ann Gibbons

ARCHAEOLOGY__

Art Stirs Uproar Down Under history." The New York

Times followed the next
it was almost the archaeological equivalent with a relatively new tech' day, boldly asserting that
of finding life on Mars. Two weeks ago, what nique called thermolumi- the find held "signs of
uI could be the biggest archaeological news in nescence (TL). "Unbe- artists who predate Homo
uJ
CD
decades erupted from the northern Austra- lievable," says archaeolo- sapiens." The authors of
lian outback, with reports of an ancient site gist John Beaton of the the forthcoming paper-
that puts humans on the continent between University of California, = ~ ~~~~~~~~~~~~~~~~ - anthropologist Richard
116,000 and 176,000 years ago-up to three Davis. "These dates are Circling in on the dream time. An- Fullagar of the Austra-
times as far back as most previous estimates. wildly out of line with ev-cient dates for Australian rock art have lian Museum, and Lesley
And the team of Australian researchers also erything else we know." sparked a furor among anthropologists. Head and David Price of
found examples of rock art-small circular Even archaeologist Rhys the School of Geosciences
Qi
carvings-that they dated to about 60,000 Jones of the Australian National University at the University of Wollongong in New
years old, more than 20,000 years older thanin Canberra, who last year made waves by South Wales-have been fielding a hail of
the most ancient known art of this kind. dating two other Australian sites to 60,000 phone calls, faxes, and e-mails ever since.
Because most anthropologists believe that years with the same method (Science, 31 The paper describes a rocky site in the
modern humans did not leave Africa until March 1995, p. 1908), warns that until more Kimberly area of northwestern Australia that
100,000 years ago, these dates, if confirmed, tests have been done, "we do not know how is known to the local aborigines as Jinmium.
would force a massive revision of human his- valid the present TL claims are." There, Fullagar's team found buried artifacts,
tory. "If it could be demonstrated [that] people The big news came out in a rather unor- including pieces of ochre and tools made
were in Australia more than 100,000 years thodox fashion: A paper on the find was from carved rock, plus thousands of circles,
ago, we would have to rethink everything wescheduled for the December issue of the Brit- all about 3 centimeters across, carved into
thought we knew about the later phases of ish journal Antiquity. But editor Christopher boulders. The team dated the sediments with
human evolution," says Stanford University Chippindale says someone at the Australian TL, which involves measuring electrons
paleoanthropologist Richard Klein. Museum in Sydney inadvertently leaked the trapped in defects in quartz crystals; these
But that's a big if. The significance of the story, and on 21 September the Sydney Mom-electrons accumulate at a regular rate but are
site hinges on its age, and many scientists are ing Herald trumpeted the discovery of an "bleached" out of the crystal by sunlight, so
skeptical about the dating, which was done "outback Stonehenge that will rewrite our that the "clock" starts at zero when a sample

SCIENCE * VOL. 274 * 4 OCTOBER 1996 33

This content downloaded from 157.182.150.22 on Sat, 17 Sep 2016 00:03:14 UTC
All use subject to http://about.jstor.org/terms