Upload
lyminh
View
217
Download
0
Embed Size (px)
Citation preview
Genetics, Language, and the Prehistory of
the Americas
Alan R. Rogers
April 2, 2018
1 / 29
Outline
I Three waves of migration or one?
I Paleoamerican hypothesis
I Population growth in South America
2 / 29
Joseph Greenberg Hypothesis
Greenberg advocatedcomparing many languages ata time, using a small list ofslowly-changing words, todetect deep relationships.
Divided Amerindian languagesinto 3 major groups—Amerind,Na-Dene (includingAthabascan), andEskimo-Aleut—which in hisview descend from 3 waves ofmigration into the Americas.
3 / 29
Dispute
Most linguists don’t acceptGreenberg’s method.
In their view, linguistic data arenot informative that far back inthe past.
Instead of 3 major groups, theyrecognize 150–180 independentlanguage families.
4 / 29
Raghavan et al (2015) Study
Sequenced 31 modernAmerindian genomes and 23ancient genomes. Also usedpreviously-published genomes.
SNP chip genotypes from 79individuals from 28populations.
5 / 29
Early North Am. similar to modern South Am.
Next few slides compare DNA of ancient fossils with that ofmodern Amerindians.
General pattern: early fossils are genetically similar to modernAmerindians farther south.
Later fossils are similar to modern Amerindians of their ownregion.
6 / 29
Pleistocene Anzick-1 fossil most similar to S
Americans
Anzick fossil, ofClovis culture,was found inMontana.
DNA most similar to S American populations.
7 / 29
Early Holocene Kennewick fossil
∼8.5 ky old,fromWashingtonstate
DNA similar toAmerindiansthroughout USand S America.
Not more similar to Amerindians of its own region, the PacificNorthwest.
8 / 29
A 6000 year old fossil
DNA still similar to Amerindians throughout US and SAmerica.
9 / 29
A 5500 year old S American fossil
More similar to Amerindians of S America.
10 / 29
A 3500 year old S American fossil
Most similar to Amerindians of its own region of S America.
11 / 29
A ∼500 year old Mexican fossil
Most similar to Amerindians of Mexico.
12 / 29
A ∼400 year old fossil from US Northeast
Most similar to Amerindians of US SW and Mexico.
An exception to the general rule.
13 / 29
A ∼200 year old fossil from Tierra del Fuego
Most similar to Amerindians of Tierra del Fuego
14 / 29
Population Relationships
Consistent with Greenberg hypothesis; Anzik & Karitiana(Amerind) are related to Athabascan (Na Dene) and (lessclosely) to Inuit and Saqqaq (Eskimo-Aleut).
15 / 29
Dating the Asian-Amerindian Split
Comparing two Amerindianpopulations with the Koryak ofeastern Siberia.
Upper panel: Karitiana are a SAmerican population thatspeaks an Amerind language(in Greenberg’s classification).
Lower panel: Athabascan ispart of Greenberg’s Na-Denelanguage family.
16 / 29
Split date contradicts GreenbergVertical axis estimatescumulative frequency ofcoalescent events.
Red: estimated from geneticdata
Blue: best-fitting modelwithout migration
Red: best-fitting model withmigration
Amerinds and Athabascansboth separated from Asiansabout 20 kya.
17 / 29
Separation wasn’t sudden
TB generations ago, theancestral population changessize.
TDIV generations ago, it splits intwo, but the subdivisionscontinue exchanging migrants.
TM generations ago, migrationstops.
18 / 29
Colonization of the Americas
19 / 29
Outline
◦ Three waves of migration or one?
I Paleoamerican hypothesis
I Population growth in South America
20 / 29
Paleoamerican Model (Gonzalez et al 2008)
I Earliest American fossils differ in skull shape from laterones.
I More like modern Australians and Melanesians.
I These earliest Americans were mostly replaced byAmerindians who arrived later.
I Evidence is from early skeletons and a few historical relictpopulations, now extinct: the Pericues and theFuego-Patagonians.
I Hypothesis predicts these populations should have DNAlike Australians and Melanesians, not Amerindians.
21 / 29
Testing the Paleoamerican modelPrincipalcomponents plot.
ModernAmerindians arethe blurry bluepoints in theupper left.
AncientAmerindians,including Pericuesand Fuego-Patagonians,
plot with modern Amerindians. No ancient Amerindians withOceanians.
22 / 29
Testing the Paleoamerican model
Pericues & Fuego-Patagonians have DNA like Amerindians,not like Oceanians.
Refutes Paleoamerican hypothesis. (Raghavan et al 2015)
23 / 29
Outline
◦ Three waves of migration or one?
◦ Paleoamerican hypothesis
I Population growth in South America
24 / 29
South American archaeological sites
25 / 29
Density of radiocarbon dates (Goldberg et al 2016)
Radiocarbon dates serve as proxy for population size.
No information in the Amazon basin.
Coastal population growth accelerates after 5 kya.
26 / 29
Temperature and population
a, temperature; b, radiocarbondensity; c, frequency of sites.
Early growth; crash at 11 kya;plateau during 10–6 kya; thenmore growth.
Megafauna went extinct 11 kya;sedentary agriculture began5.5–3.5 kya.
1000-year oscillations duringplateau. Don’t know why.
(Goldberg et al 2016)
27 / 29
Fitting a model
Top: growth rate;Bottom: population size.
Before 6 kya, growthrate declines aspopulation increases.
After 6 kya, growth ratestable as populationincreases.
(Goldberg et al 2016)
28 / 29
Summary
I All Amerindians, including ancient ones, separated fromAsians about 22 kya.
I Separation wasn’t sudden.
I Na-Dene (incl Athabascans) separated from Amerindsabout 13 kya.
I Regional genetic differences developed gradually, overthousands of years.
I Paleoamerican hypothesis is false: earliest Americans notrelated to Oceanians.
I Population history of South American: initial growthleading to a plateau, then renewed growth afteragriculture.
29 / 29