Dispatch R925
Human evolution: The southern route to Asia
Todd R. Disotell
Research on human origins has tended to focus on the
origins of western Eurasians; only recently have genetic
studies examined south and east Asian populations in
depth. Recent work suggests that the supposed Aryan
invasion of India 3,0004,000 years ago was much less
significant than is generally believed.
Address: Department of Anthropology, New York University,
25 Waverly Place, New York, New York 10003, USA.
Current Biology 1999, 9:R925R928
0960-9822/99/$ see front matter
1999 Elsevier Science Ltd. All rights reserved.
According to the multiregional model of human evolu-
tion, archaic humans originated in Africa, migrated
throughout the Old World over a million years ago, and
then evolved into modern form multiple times in differ-
ent areas of the Old World, with enough gene flow
between these regions to prevent speciation [1]. In this
scenario, east Asians, Australians, Europeans and
Africans would each have had relatively ancient separate
ancestries. This theory is supported predominately by
paleoanthropologists who see, for instance, that the
hominid fossils from Australasia (Indonesia, New Guinea
and Australia) show a continuous anatomic sequencing
during the Pleistocene that is un-interrupted by African
migrants at any time [1]. Southeast and south Asian
populations are also often thought to be derived from
the admixture of various combinations of western
Eurasians (Caucasoids), east Asians and Australasians.
The combinations of phenotypic traits in some of these
populations could then be viewed either as the results of
such admixture, or as traits selected for a particular envi-
ronment (for example dark skin).
The widely supported recent replacement model, on the
other hand, posits a relatively recent African origin for all
modern humans, with a subsequent dispersal throughout
the Old World that completely replaced the existing
archaic population (reviewed in [2]). A model with a
single migration out of Africa, however, is receiving less
support as additional fossil and genetic studies more
fully characterize human diversity, both past and
present. Several researchers [3,4] have proposed that two
geographical routes were taken by early modern
migrants from Africa: a northern route through north
Africa and the Middle East towards western Eurasia, and
a southern route through Ethiopia and the Arabian
peninsula towards South Asia. A study recently pub-
lished in Current Biology [5] provides important new
support for the southern route hypothesis, from the
analysis of a large sample of south Asian mitochondrial
(mt)DNA sequences and recent discoveries of previ-
ously unknown mtDNA types in Ethiopia.
The initial mtDNA study of Cann et al. [6] and subsequent
studies have found far greater genetic variation within
Africa than in the rest of the Old World, consistent with
the view that the African population as a whole originated
between 100,000 and 200,000 years ago, and that the dis-
persal outside of Africa occurred much less than 100,000
years ago [7,8]. Population differentiation thus probably
occurred for a considerable period within Africa before
populations dispersed elsewhere into the Old World [9].
One of the weaknesses of many genetic studies of modern
human origins is the difficulty of both carrying out and
interpreting phylogenetic analyses. The most widely used
technique for inferring evolutionary relationships from
genetic sequence data is maximum parsimony. This
technique calculates the number of evolutionary events
nucleotide substitutions in the case of DNA sequences
over all possible bifurcating trees linking the samples
being analyzed. As many tree topologies may require the
same number of events, multiple equally parsimonious
trees are often found. This is especially true for large data
sets or those in which only a few differences exist between
the various samples. If some of the changes shared
between individuals have arisen independently known
as homoplasies maximum parsimony approaches may
give misleading results.
An alternative approach to inferring phylogenies from
intraspecific data, which often consist of large samples
with small distances between individuals, is the median-
joining technique recently put forth by Bandelt et al. [10].
This method creates a network out of data for non-recom-
bining parts of the genome, such as mtDNA or Y chromo-
some sequences, by joining individuals who differ by only
a specified number of changes into clusters, which them-
selves are linked to other clusters and the network as a
whole (Figure 1). This method can tolerate a reasonable
amount of homoplasy. Instead of finding tens of thousands
of equally parsimonious trees with little resolution, the
median-network approach will produce a single network
with alternative potential evolutionary paths between
individuals and clusters of individuals [10] (Figure 1). It is
this method that Kivisild et al. [5] have used to effect in
their study of the origins of Indian populations.
While a great deal of research effort has focused on the
origins of Europeans and the fate of the neanderthals,
R926 Current Biology Vol 9 No 24
Figure 1
H6
H11
H13
H10
H14
H12
H1
H2
H3
H4
H5
H6
H7
H8
H9
H14
H9
H5
Current Biology
surprisingly few studies have examined large samples
from Asian and Australian populations. For instance, the
patterns of genetic diversity in India soon to be home
to the worlds largest population have only recently
come under scrutiny using the latest molecular and ana-
lytical techniques. Kivisild et al. [5] have now reported a
study that includes an analysis of 550 Indian mtDNA
samples using the median-joining approach. All of the
Indian mtDNA types found could be derived from the
African mtDNA haplogroup L3a, which is of course con-
sistent with an African origin, and if the timing is right,
with the recent replacement hypothesis. (A haplogroup
is a cluster of mtDNA types consisting of closely related
sequences that differ by only a few nucleotide substitu-
tions.) They found that 60% of Indian mtDNA types
belong to the Asian-specific haplogroup M.
Kivisild et al.s [5] large scale study also found that the
mtDNA haplogroup U, until now thought to be a
western Eurasian marker based on much smaller studies,
is the second most frequent type in India, with a 20%
frequency. This haplogroup is actually composed of
seven subtypes: the western Eurasian subtypes are
found at a low frequency in India, and vice versa. The
coalescence, or time of origin, for the western Eurasian
and Indian U2 subtypes was calculated as 53,000 years
ago. Another haplogroup, U7, found at much higher fre-
quencies in India and rarely in western Eurasia, has an
estimated coalescence date of 32,000 years ago.
Where western Eurasian mtDNA types are found among
the Indian sample, their frequencies are more than ten
times lower than in western Eurasia. Within these shared
types, the divergence times between Indian and western
Asian types estimated from the minimal distances
between their corresponding mtDNA hypervariable
On the left is shown a strict consensus tree of
H3 the numerous equally parsimonious trees
H12
linking 14 mtDNA types from a sample of
61 humans [16]. On the right is shown the
medianjoining network linking the 14 mtDNA
H1 H10 types with one change between each node
(after [10]). Note type H14 could be derived
from either H1 or H8 and type H10 from
H2 either H1 or H7. Additional information,
H8 H7
based on geography and from other studies,
suggests that the links represented by the
grey lines are less plausible [10].
H4
H11
H13
region sequences and a reasonable mutation rate is on
the order of 9,000 years ago, which Kivisild et al. [5]
interpret as indicative of a small amount of admixture,
possibly due to the expansion of agricultural populations
from the fertile crescent. These findings, coupled with the
recently discovered presence of haplogroup U in Ethiopia
[11], support a scenario in which a northeast African popu-
lation dispersed out of Africa into India, presumably
through the Arabian peninsula, before 50,000 years ago
(Figure 2). Other migrations into India also occurred, but
rarely from western Eurasian populations.
The supposed Aryan invasion of India 3,0004,000 years
before present therefore did not make a major splash in
the Indian gene pool. This is especially counter-indi-
cated by the presence of equal, though very low, fre-
quencies of the western Eurasian mtDNA types in both
southern and northern India. Thus, the caucasoid fea-
tures of south Asians may best be considered pre-cauca-
soid that is, part of a diverse north or north east
African gene pool that yielded separate origins for
western Eurasian and southern Asian populations over
50,000 years ago.
Recent large scale genetic studies of east Asian and Aus-
tralasian populations are consistent with this scenario. Chu
et al. [12] typed from between 15 and 30 microsatellites in
28 Chinese populations as part of the Chinese Human
Genome Diversity Project. They inferred that the genetic
data do not support an independent origin of modern
humans in China, as proposed by the multiregional model,
but rather that the ancestors of the Chinese dispersed
from south eastern Asia. Thus, rather than being an
ancient population that intermixed with other populations
at its periphery, especially to the south, the far east Asian
population may be relatively recently derived from south

Figure 2
The northern (green) and southern (red)
human dispersal routes inferred from patterns
of mtDNA variation. According to Kivisild et al.
[5], the more recent 9,000 year old dispersal
from the fertile crescent region into southern
and eastern Asia (blue) provided a very minor
contribution to Asian mtDNA gene pools (see 53,000
years
text for details).
53,000
years
Current Biology
east Asian ancestors, who themselves are derived from
populations dispersing from the Indian subcontinent just
over 50,000 years ago (Figure 2).
One of the more perplexing questions about modern
human evolution centers around the origins of the aborigi-
nal Australians and other people of the Sahul the
Pleistocene landmass that connected Australia, New
Guinea, Tasmania and many of the islands in the region.
Archeological evidence suggests that Australia was
colonized between 40,000 and 60,000 years ago [13]. If the
multiregional model is ruled out, and the million or so year
old archaic inhabitants of the region did not leave their
genes behind, as most molecular evidence suggests, where
did these populations come from? Conflicting hypotheses
have been proposed based on multiple genetic data sets
and types of analysis. Multiple migrations and dispersals
probably further confound these analyses.
Some recent mtDNA-based studies (for example [14])
link Australian aborigines with populations from New
Guinea, though perhaps with several migration events
leading to distinct subgroups on each island. Other
studies have found a link between the Australians, but not
New Guineans, and Indian populations [15]. Hypervari-
able sites in the most commonly sequenced region of the
mitochondrial genome make the application of tree-based
comparisons to global samples particularly difficult, even
for median-joining approaches [14]. Estimating coales-
cence dates is also problematic for these populations,
perhaps because of increased genetic drift as a result of
their relative isolation. Despite the confusion as to from
where and when Sahulian populations are derived
south Asia, south east Asia or perhaps both a scenario
in which they are ultimately derived from an African pop-
Dispatch R927
9,000
years
>30,000
years
>40,000
years
ulation who dispersed by the southern route within the
last 60,000 years is becoming increasingly likely (Figure 2).
This scenario could also explain one of the enduring
mysteries of human morphological variation. South Asian
populations are typically classified as caucasoid, despite
numerous phenotypic features that resemble Africans
and Australian aborigines. These may be ancestrally
retained pre-caucasoid traits derived from a north or
north-east African population before the western
Eurasian/south and east Asian divergence. The dark-
skinned Australian aborigines, who often cluster with
some African populations when morphometric compar-
isons of skulls are made, may also share many ancestral
traits with the original founding population. Various
Indian Ocean and Pacific island populations often
display a constellation of negrito traits, including small
stature, dark skin and tightly curled hair. Usually, these
traits are explained as evolving due to a combination of
mutation, isolation, drift, and selection to tropical envi-
ronments over hundreds of thousands of years. If the
southern route scenario is correct, these traits may be
the results of selection of a quite variable population that
expanded along the southern periphery of the Asian con-
tinent relatively rapidly between 50,000 and 60,000 years
ago. Rather than being perplexed by these features, it is
clear that they need to be reevaluated in light of the
evidence supporting such a scenario.
References
1. Thorne AG, Wolpoff MH: The multiregional evolution of humans.
Sci Am 1992, 266:76-83.
2. Disotell TR: Origins of modern humans still look recent. Curr Biol
1999, 9:R647-R650.
3. Cavalli-Sforza LL, Menozzi P, Piazza A. The History and Geography
of Human Genes. Princeton, New Jersey: Princeton University
Press; 1994.
R928 Current Biology Vol 9 No 24

4. Lahr MM, Foley R: Multiple dispersals and modern human origins.

Evol Anthrop 1994, 3:48-60.
5. Kivisild T, Bamshad MJ, Kaldma K, Metspalu M, Metspalu E, Reidla M,
Laos S, Parik J, Watkins WS, Dixon ME, et al.: Deep common
If you found this dispatch interesting, you might also want
ancestry of Indian and western Eurasian mtDNA lineages. Curr to read the December 1999 issue of
Biol 1999, 9:1331-1334.
6. Cann RL, Stoneking M, Wilson AC: Mitochondrial DNA and human
evolution. Nature 1987, 325:31-36.
Current Opinion in
7. Watson E, Forster P, Richards M, Bandelt H-J: Mitochondrial
footprints of human expansions in Africa. Am J Hum Genet 1997, Genetics & Development
61:691-704.
8. Mountain JL, Hebert JM, Bhattacharyya S, Underhill PA, Ottolenghi C, which included the following reviews, edited
Gadgil M, Cavalli-Sforza LL: Demographic history of India and
mtDNA-sequence diversity. Am J Hum Genet 1995, 56:979-992. by Phil Green and Eugene Koonin, on
9. Harris EE, Hey J: X chromosome evidence for ancient human Genomes and evolution:
histories. Proc Natl Acad Sci USA 1999, 96:3320-3324.
10. Bandelt H-J, Forster P, Rhl A: Median-joining networks for
Comparative genomics and evolutionary biology
inferring intraspecific phylogenies. Mol Biol Evol 1999, 16:37-48.
11. Passarino G, Semino O, Quintana-Murci L, Excoffier L, Hammer M, Alexy S Kondrashov
Santachiara-Benerecetti AS: Different genetic components in the
Ethiopian population, identified by mtDNA and Y-chromosome
Orthologs, paralogs and genome comparisons
polymorphisms. Am J Hum Genet 1998, 62:420-434. J Peter Gogarten and Lorraine Olendzenski
12. Chu JY, Huang W, Kuang SQ, Wang JM, Xu JJ, Chu ZT, Yang ZQ,
Lin KQ, Li P, Wu M, et al.: Genetic relationship of populations in
Coding sequence evolution
China. Proc Natl Acad Sci USA 1999, 95:11763-11768. Martin Kreitman and Josep M Comeron
13. Klein RG: The Human Career: Human Biological and Culture Origins
2nd ed. Chicago: University of Chicago Press, 1999.
Selfish operons: the evolutionary impact of gene
14. van Holst Pellekaan SM, Frommer M, Sved JA, Boettcher B: clustering in prokaryotes and eukaryotes
Mitochondrial control-region sequence variation in Aboriginal Jeffrey Lawrence
Australians. Am J Hum Genet 1998, 62:435-449.
15. Redd AJ, Stoneking M: Peopling of Sahul: mtDNA variation in Shaping the genome - restriction-modification
Aboriginal Australian and Papua New Guinean populations. Am systems as mobile genetic elements
J Hum Genet 1999, 65:808-828. Ichizo Kobayashi, Ayaka Nobusato, Noriko Kobayashi-
16. Nachman MW, Brown WM, Stoneking M, Aquadro CF: Nonneutral Takahashi and Ikuo Uchiyama
mitochondrial DNA variation in humans and chimpanzees.
Genetics 1996, 142:953-963. Interspersed repeats and other mementos of
transposable elements in mammalian genomes
Arian FA Smit
Insights into the evolutionary process of genome
degradation
Jan O Andersson and Siv GE Andersson
The nature of the last universal common ancestor
David Penny and Anthony Poole
Evolution of organellar genomes
Michael W Gray
Origin of multicellular eukaryotes - insights from
proteome comparisons
L Aravind and G Subramanian
Noncoding RNA genes
Sean R Eddy
Contribution of genomics to bacterial pathogenesis
Dawn Field, Derek Hood and Richard Moxon
The genome of the malaria parasite
Malcolm J Gardner
The minimal genome concept
Arcady Mushegian
Observing the living genome
Tracy L Ferea and Patrick O Brown

The full text of Current Opinion in Genetics &

Development is in the BioMedNet library at
http://BioMedNet.com/cbiology/gen