The evolving story of rice evolution
Duncan A. Vaughan
*, Bao-Rong Lu
, Norihiko Tomooka
National Institute of Agrobiological Sciences, Tsukuba, Japan
Ministry of Education Key Laboratory for Biodiversity Science and Ecological Engineering,
Institute of Biodiversity Science, Fudan University, Shanghai 200433, China
Received 30 October 2007; received in revised form 24 January 2008; accepted 24 January 2008
Available online 13 February 2008
Recent research related to evolution in the primary gene pool of rice, which consists of Oryza species with the A-genome, provides new
perspectives related to current and past eco-genetic setting of rice and its wild relatives and fresh insights into rice domestication. In Asia the traits
of the rice domestication syndrome are many but due to the remarkable diversiﬁcation of rice and introgression with wild rice, few traits are
consistently different between wild and domesticated rice. Reduced shattering and reduced dormancy are the principal traits of domestication in
rice. Using the principal criteria for distinguishing single and multiple origins of crops, recent key research results do not support a polyphyletic
origin of domesticated rice in distinctly different geographic regions. While domestication is a long-term process and continues today, a single
event during domestication, the selection of the non-shattering sh4 allele, resulted in rice becoming a species dependent on humans for survival -
domesticated. Here the apparent contradictions between a single origin of Asian rice and deep genetic divisions seen in rice germplasm are resolved
based on a hypothesis of cycles of introgression, selection and diversiﬁcation from non-shattering domesticated rice, importantly in the initial
stages in its center of origin in the region of the Yangtze river valley, and subsequently beyond, as domesticated rice spread. The evolution of
African rice differs from Asian rice mainly in the more restricted gene pool of wild rice from which it was domesticated, ecological diversiﬁcation
rather than eco-geographic diversiﬁcation, and historic introgression from the Asian rice gene pool. The genetics of post-domestication evolutionin
Asian rice is well illustrated by changes at the waxy locus. For both Asian and African rice becoming domesticated was a single event, it was the
subsequent evolution that led to their genetic complexity.
#2008 Elsevier Ireland Ltd. All rights reserved.
Keywords: Wild rice; Domestication; Genetic resources; Oryza; Waxy gene
1. Introduction . . . .............................................................................. 395
2. Setting the scene—the primary gene pool of rice and its components . ........................................ 395
3. Before domestication—the A-genome Oryza gene pool evolution ............................................ 398
4. Domestication . . .............................................................................. 398
4.1. The domestication syndrome ................................................................. 398
4.2. Single or multiple origins of Asian rice ......................................................... 399
4.2.1. Founder effects . . . ................................................................. 400
4.2.2. Domestication traits ................................................................. 400
4.2.3. Species diversity . . ................................................................. 400
4.3. The domestication of Asian rice. . ............................................................. 400
4.3.1. Diversity of wild and cultivated rice in China . . . ............................................ 400
4.3.2. An eco-genetic hypothesis for the domestication of Asian rice.................................... 401
4.4. Evolution of rice in Africa . ................................................................. 403
5. Post-domestication—the waxy locus................................................................. 404
vailable online at www.sciencedirect.com
Plant Science 174 (2008) 394–408
This review was based on literature available to the authors up to September 2007.
* Corresponding author. Tel.: +81 298 38 7474; fax: +81 298 38 7408.
E-mail address: email@example.com (D.A. Vaughan).
0168-9452/$ – see front matter #2008 Elsevier Ireland Ltd. All rights reserved.
6. Conclusions. . ................................................................................ 405
Acknowledgements ............................................................................ 406
References . . ................................................................................ 406
The sequencing of the whole genome of rice, more than
once, has predictably resulted in a rapid surge in research on
rice genetics and evolution. Cloning rice genes has concen-
trated on the most important domestication related genes. Two
recent papers describe cloning different genes related to
shattering in rice [1,2]. Two other papers have subsequently
appeared related to evolution of shattering in rice [3,4]. Aside
from molecular studies, new studies in other branches of plant
science, such as archaeobotany, are furnishing information that
are providing new perspectives on the evolution of rice.
It might be expected that recent rice research would resolve
questions concerning the evolution of rice, but apparently
conﬂicting results and their interpretations have yet to resolve
some central questions. The objective of this paper is to discuss
topics related to the evolution of rice that have recently
provided new insights. We ﬁrst discuss the primary gene pool of
the cultivated rices before discussing research related to the
domestication of rice in Asia and Africa.
2. Setting the scene—the primary gene pool of rice and
To understand the domestication of Asian and African rice it
is necessary to understand the gene pools from which they
came. That requires understanding these gene pools as they are
found today and also how they might have been in the past.
The primary gene pool of a crop is considered to be
equivalent to a biological species consisting of germplasm that
can be crossed resulting in fertile hybrids . There are many
barriers to hybridisation in the A-genome Oryza (for review see
), but hybridisation does occur among these species when
they grow together and ﬂower at the same time. In Asia
introgression between the various A-genome Oryza species is
common [7,8]. In Africa introgression and hybrids between
introduced Oryza sativa and wild and cultivated African species
(also with the A-genome) have been reported [9,10]. Weedy
forms between the Latin American A-genome wild species,
Oryza glumaepatula, and rice (O. sativa) are commonly found
in Venezuela (Dr. Zaida Lentini, CIAT, personal communica-
tion 2007). Artiﬁcial A-genome interspeciﬁc hybrids have
nearly normal meiosis but have variable pollen and seed
Two cultivated rice species were domesticated from the A-
genome of Oryza independently in Asia and Africa, but
probably by a similar processes based on contemporary ways in
which wild rice is harvested in these continents  (Figs. 1 and
2). The cultivated Asian rice (O. sativa) is now spread
worldwide, and African rice (Oryza glaberrima)isnow
conﬁned almost exclusively to West Africa. Recently a
breeding program that has focussed on combining the best
qualities of Asian and African rice has resulted in a series of
varieties called NERICA (New Rice for Africa) rices that are
spreading to many regions of Africa [14,15].
The taxonomy of the A-genome Oryza species has long been
‘a matter of opinion’, and the distinction of species has mainly
been based on three criteria: geography, annual/perennial habit
and cultivated or wild habitat (Table 1). These criteria do not
provide reliable key morphological characters for distinguish-
ing species. Introgression among A-genome Oryza species is
widespread and only Oryza longistaminata can reliably and
consistently be distinguished from other species by its well-
developed rhizomes. However, the species names provided in
Table 1 are at present the best summary of A-genome Oryza
species diversity, the most commonly found names in the
literature and most helpful categories for understanding the
primary gene pool of rice.
Fig. 1. Plants of O. nivara twisted together to prevent shattered seeds dropping
thus facilitating their harvest, Madhya Pradesh, India.
Fig. 2. Harvesting O. ruﬁpogon in West Bengal, India.
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 395
Many accessions in gene banks are not ‘pure’ representatives
of these species especially the wild species and for a variety of
reasons may be misidentiﬁed [16–18]. Many accessions of
African cultivated rice, O. glaberrima, have genes introgressed
from O. sativa . In addition, geographic origin alone is
insufﬁcient to identify a ‘species’ because not only O. sativa but
also other A-genome species have been introduced to new
geographic regions. An accession of wild rice from Cuba is an
Asian Oryza ruﬁpogon based on various analyses [19,20].O.
longistaminata has been collected from Martinique, in the
Perennial A-genome wild Oryza have considerable intra-
speciﬁc genetic variation. Three regional variants of O.
glumaepatula can be distinguished (a) the Central American,
Caribbean, and northern South American type (perennial), (b)
the Amazonian type (perennial-intermediate) and (c) the
central South American type (perennial) . Five distinct
groups of O. longistaminata are found based on ecology and
self-incompatibility systems .O. ruﬁpogon has consider-
able geographic variation with O. ruﬁpogon from New Guinea
being recognised as distinct and called ‘Oceanian’ type .
An understanding of the ecological setting of wild rice in
Asia is critical when considering domestication of rice. One of
the major questions regarding the evolution of rice is from
which type of Asian wild rice did it evolved? Perennial O.
ruﬁpogon has many ecotypes and is widely distributed across
Asia to Papua New Guinea and Australia (Fig. 3a). The
principal ecotypes of O. ruﬁpogon differ by the degree to which
they produce seeds. Asian O. ruﬁpogon tends to be poorly
represented in germplasm collections because populations in
many regions are primarily vegetative. This is especially the
case in equatorial Asia (Sumatra and Kalimantan, Indonesia)
where year round rainfall results in lakes and rivers with a
relatively stable water level. O. ruﬁpogon growing in these
conditions is mainly vegetative and seed production is low. The
second main ecotype of O. ruﬁpogon exhibits mixed mating,
with both vegetative and seed reproduction, and copiously
produces seeds. High seed producing ecotypes of O. ruﬁpogon
are found in areas where water levels rise and fall considerably
between seasons, such as in southern Papua New Guinea. In
these areas O. ruﬁpogon grows where the soil retains residual
moisture enabling populations to survive from season to season.
Such ecotypes can also be found in continental Asia, but
because of introgression from cultivated rice it is not possible to
be sure which O. ruﬁpogon are high seed producing as a result
of introgression from cultivated rice. One type of high seed
producing O. ruﬁpogon from Thailand is considered to be an
intermediate type between annual and perennial wild rice .
High seed producing wild ecotypes of perennial O. ruﬁpogon
do exist and would have been an attractive source of nutrition in
the age of hunters and gatherers. Today in India such ‘‘wild’’ O.
ruﬁpogon is harvested (Fig. 2). Natural habitats of wild O.
ruﬁpogon are generally lakes or waterways surrounded by
forest. Disturbance in these habitats is primarily the result of
The annual, inbreeding, Oryza nivara, has a more restricted
distribution than the perennial, O. ruﬁpogon (Fig. 3b). O. nivara
is most common in the area of South and Southeast Asia with
severe dry seasons. This is clearly seen in Sri Lanka where O.
nivara is almost completely conﬁned to the dry zone and O.
ruﬁpogon to the wet zone of the country. The natural habitat of
O. nivara is uncommon but can be found in National Parks of
Sri Lanka where it grows at the margins of seasonal pools in
grassland. A characteristic of such habitats is that they are more
Geographic distribution, life cycle and cultivation status of A-genome Oryza species
Annual/perennial Wild/cultivated Latin America Africa Asia Australia and
Perennial Wild O. glumaepatula Steud. O. longistaminata
A. Chev. et Roehr.
O. ruﬁpogon Griff.
Annual Wild O. barthii A. Chev. O. nivara Sharma et Shastry O. meridionalis Ng
Annual Cultivated O. glaberrima Steud. O. sativa L. (now worldwide)
Fig. 3. (a) Distribution of O. ruﬁpogon and (b) distribution of O. nivara (updated from ).
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408396
disturbed than habitats of perennial O. ruﬁpogon, because these
pools of water, in seasonally dry areas with few water sources,
are frequented daily by a variety of animals. Thus, O. nivara is
adapted to frequent grazing that requires rapid seed production
for survival, and trampling that results in compacted soil in
which seeds survive and germinate.
Introgression from cultivated rice to wild rice is a common
occurrence where they are sympatric [6,7,25]. Due to ﬂoral
structure perennial outcrossing wild rice is more likely to be a
pollen receiver than annual inbreeding wild rice . During
cultivation of rice, selection pressure would have favoured high
seed producing populations and hence inbreeding. Subse-
quently when rice was fully domesticated (non-shattering) it
would have been sympatric with wild rice as it spread.
Hybridisation and introgression would have occurred in a
similar way to that recorded today. However, it is not known the
degree of inbreeding in early domesticated rice. Compared to
today higher outcrossing and introgression may have been a
feature of early domesticated rice when there would have been
few barriers to hybridisation among different types of wild,
cultivated and newly domesticated rice. Introgression from
wild rice of different genetic backgrounds into newly
domesticated rice could readily have occurred enabling
domesticated rice to adapt to new environments.
Most of the germplasm in gene banks of Asian wild rice was
collected from human-disturbed or human-made habitats such
as roadside ditches and irrigation channels. These are areas
where introgression may have occurred between wild rice and
rice. Consequently conclusions regarding the genetics of gene
bank germplasm should be understood in this context. While
high seed producing O. ruﬁpogon and O. nivara are both
adapted to disturbed habitats, they differ in the degree of
disturbance and degree to which residual moisture occurs in the
soil. Thus, the presence of annual O.nivara and perennial O.
ruﬁpogon in Asia are an adaptive response to habitat conditions
with some genetic isolation due to differences in ﬂowering time
and also geographic variation reﬂecting isolation by distance.
The differences in ﬂowering time can result in local scale
variation while geographic isolation is seen at the regional level
and revealed by quantitative trait analysis . Sequencing of
particular genes  and multiple neutral genes  for
polymorphisms does not reveal differences between O. nivara
and O. ruﬁpogon. This may reﬂect recent divergence and/or
continual gene ﬂow between these two ecotypes in mainland
Asian O. sativa has two major varietal groups or subspecies,
japonica (keng) and indica (hsien), that have traditionally been
recognised in the Chinese language . Analyses of
germplasm collections have revealed the main components
of Asian rice diversity at the end of the 20th Century. Six groups
of rice varieties were recognised using isozymes—indica, aus
(early summer), ashwina (early deep water), rayada (long
duration deep water), aromatic and japonica . Subsequently
the importance of the aus (indica aligned) and aromatic
(japonica aligned) varieties has been conﬁrmed using SSR
markers. Differences were also recognised between temperate
and tropical japonica varieties . Within the tropical
japonica varieties are a morphologically distinct group from
Indonesia called bulu or javanica that have few tillers, long
panicles and seeds with awns. The other two varietal groups
recognised by isozyme analysis, ashwina and rayada, were
selected for the environments of some villages in the Ganges
river delta. Ashwina are early maturing deepwater varieties
lacking strong photoperiod sensitivity while rayada are long
duration, deepwater varieties lacking secondary dormancy. The
unusual characteristics of these two small varietial groups may
provide insights into rice domestication in the complex
environments of deepwater areas. Rice diversity studies have
relied on the germplasm collected for the past 40 years. The
traditional lowland varieties of some areas were not or only
poorly collected prior to the introduction of improved varieties
from breeding programs, for example some parts of Myanmar.
Thus, studies of rice genetic diversity represent a biased
sampling of germplasm from a restricted time period in rice
evolution and important germplasm for understanding rice
evolution undoubtedly has been lost.
West African ecological diversiﬁcation of O. glaberrima is
similar to that of O. sativa in Asia but O. glaberrima has less
genetic diversity. In O. glaberrima some distinct variations
found using SSR markers are thought to be associated with
ecological differentiation—ﬂoating, non-ﬂoating lowland and
upland rice . Sophisticated strategies for rice production
have been developed within West African societies. Despite O.
glaberrima having less diversity than O. sativa, the parallels are
striking in the way African and Asian societies have selected
rice and manipulated the environment to suit the crop. For
example, the tidal wetlands rice cultivation practiced by the
Diola of Senegal  is very similar to that described early in
the 20th Century of ‘kaipal’ cultivation in Kerala, India .
People in both areas independently developed complex
processes of diking, desalination, ridging and transplanting
to cope with seasonal seawater intrusion. In East Asia land
inundated with seawater was reclaimed for rice cultivation
using somewhat different approaches such as the initial use of
salt tolerant plants, like barnyard millet (Echinochloa crus-
galli) to improve the land .
In contrast to many other crops, Asian and African rice is
primarily grown where it evolved and, hence, where its wild
relatives occur. In Asia, introgression from wild rice to
cultivated rice is more likely to occur in indica varieties than
japonica varieties because of where these two varietal groups
are grown. Tropical japonica varieties in mainland Asia are
generally grown in highland where wild rice does not grow and
in insular Asia, particularly Java, where A-genome wild rice is
rarely in proximity to rice ﬁelds. Indica varieties are mainly
lowland varieties that on mainland Asia often grow sympatric
with wild rice.
Finally, a major and increasingly important component of
the primary gene pool of rice is weedy rice. Weedy rice grows in
rice ﬁelds and is adapted to the rice cropping system. Due to its
genetic similarity to rice it is difﬁcult to eliminate from rice
ﬁelds without careful attention to land preparation. As
discussed above introgression between components of the
primary gene pool can result in weedy rice derived from
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 397
hybridisation events. In addition, weedy rice can be derived
from natural selection for shattering in rice growing environ-
ments due to cultural practices. Direct seeding, particularly
where land preparation is inadequate, can quickly result in
shattering and unwanted non-shattering off types because rice
plants are scattered, rather than in transplanted rows, making
weeding difﬁcult. A full discussion of this topic can be found in
six chapters of .
3. Before domestication—the A-genome Oryza gene
Oryza is an ancient genus allied to the bamboos . Based
on phytolith evidence the ancestors of Oryza were present 65
million years ago (Mya) . Molecular clock data suggests
that diversiﬁcation of the genus Oryza occurred 8–14 Mya 
and diversiﬁcation within the A-genome over the past 2 Mya
Given the current pan-tropical distribution of A-genome
Oryza many scientists have tried to elucidate the evolutionary
relationships within and between these species. Some scientists
consider O. longistaminata is the most diverged of all the A-
genome Oryza species .O. longistaminata is a rhizoma-
tous, perennial, self-incompatible species found in much of
sub-Saharan Africa. Many molecular studies using RAPD ,
RFLP [17,19], MITE-AFLP , and intron sequences of four
nuclear genes  point to Oryza meridionalis as the most
diverged of the A-genome Oryza species. This inbreeding,
annual species, found in habitats similar to O. nivara in
mainland Asia, is distributed across northern Australia and has
been found in Irian Jaya, Indonesia.
If O. meridionalis is closest to the ancestral type of the A-
genome, Oryza germplasm collections are missing an out-
crossing, perennial species linked to the ancestor of A-genome
Oryza, since all Oryza closely related to A-genome species are
perennial. Comprehensive surveys of Oryza in northern
Australia and New Guinea are lacking, so it is possible there
is a perennial type similar to O. meridionalis. A perennial O.
ruﬁpogon ecotype was found in the Sepik River of northern
Papua New Guinea and it has some genome regions
characteristic of O. meridionalis (, Hai Fei Zhou et al.
Institute of Botany, Chinese Academy of Sciences, unpublished
data).We still have incomplete knowledge about the A-genome
Oryza from Australia and New Guinea, so further collecting
and analysis of germplasm from that region should be a priority.
Given that O. meridionalis and O. longistaminata are the
most diverged of the A-genome Oryza species, the question
arises how were these species, conﬁned to different continents,
were dispersed? The Gondwanaland hypothesis for the
distribution of Oryza [43,44] is no longer tenable based on
what is now known about the time grasses and Oryza arose and
when continental drift occurred. A hypothesis proposing that
animals, including birds and humans, are the major factor in
disseminating Oryza species was put forth . In addition to
Oryza, species in the genus Sorghum,Gossypium and Vigna,
among others, are also found in Africa and Australia. Further
understanding of the relationships between O. meridionalis and
O. longistaminata, that have differences in life cycle, breeding
and morphological characteristics, are needed to gain a better
understanding the origins of the A-genome.
4.1. The domestication syndrome
Among crops Asian rice varietal diversity is remarkable.
This is a reﬂection of its early domestication and subsequent
spread to one of the worlds most geographically diverse
regions around Yunnan province, China, introgression from
wild rice, and selection for a broad range of ecological
conditions and taste preferences by many ethnic groups
[46,47]. In addition, the selection associated with domestica-
tion resulted in a greater degree of inbreeding and reduced
recombination. Thus, both beneﬁcial and deleterious muta-
tions, accumulated in diverse lines . A consequence of the
high level of rice diversity is that there are few domestication
related traits that are consistently associated with cultivated
rice, except for two.
These two traits were key to the domestication of rice—
degree of spikelet shattering and loss of strong secondary
dormancy. In rice, shattering and dormancy are complex traits
with many QTLs for each trait in many linkage groups of the
rice genome . Both traits vary considerably depending on
the variety or varietal group. Ease of shattering differs among
different types of rice. Japonica varieties do not have an
abscission layer at the spikelet (seed) base, indica varieties have
a partial abscission layer while wild and weedy rice have a
complete abscission layer. Differences in the ease of shattering
is reﬂected in cultural practices associated with traditional
methods of threshing.
Extreme range for three characters in rice
Category Species name Panicle length Grain length Grain width
Longest (cm) Shortest (cm) Longest (mm) Shortest (mm) Widest (mm) Narrowest (mm)
Cultivated O. sativa 43 10 14.4 3.5 4.9 1.4
O. glaberrima 32 14 10.5 6.1 3.6 2.2
Wild O. barthii 40 9 12.3 7.7 3.5 2.2
O. nivara 35 9 10.5 5.2 3.4 1.5
O. ruﬁpogon 40 12 11 4.8 3.2 1.1
Based on data from the T.T. Chang Genetic Resources Center, IRRI, rice collection, June 2007.
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408398
Dormancy is a complex trait consisting of seed and hull
dormancy with QTL for dormancy found on most of the
chromosomes of rice . Secondary dormancy remains an
important characteristic in indica varieties, where seeds retain
viability during hot, humid conditions between planting
seasons. The most detailed studies of dormancy have been
conducted with an accession of wild-like weedy rice from
Thailand (accession SS18-2) that exhibits hull-imposed
dormancy [50–55]. One dormancy related QTL in this
accession on chromosome 7, explaining about 13% of the
phenotypic variance, is closely linked (or pleiotropic) to the
gene for red grain and some other domestication related QTL
[50–53]. But the most important dormancy related QTL in this
Thai accession was located on chromosome 12 explaining
about 50% of the phenotypic variance [51,53]. This chromo-
some 12 QTL may be the most important to study in relation to
rice domestication not only because of its large effect but also
unlike some other dormancy QTL it is largely independent of
loci for interrelated characteristics such as red pericarp, hull
colour and awn length . If this major dormancy QTL on
chromosome 12 is widespread in the rice genepool, when it is
cloned , it may be as informative as the shattering QTL sh4
has been regarding rice domestication.
Domestication of many crops is associated with an increase
in size of the organs that are consumed. Thus, cultivated
sorghum has much larger seed size than wild sorghum .
However, rice has not always been selected for larger seed size
(Table 2). Most rice varieties have a seed size little different
from wild rice. The International Rice Research Institute rice
collection of cultivated rice has accessions up to 30% longer
and shorter than wild rice. In the case of African rice, Oryza
barthii accessions have the longest and shortest seeds. They
exceed those of the longest and shortest seeds of the cultigen, O.
glaberrima by 15 and 20%, respectively (Table 2). Thus, seed
size is not a reliable character to distinguish wild and
domesticated rice. Other traits have been more important
during domestication such as number of seeds per panicle and
synchronous maturity. The number of panicles per plant is
highly variable among rice varieties. Some varietal groups,
particularly tropical japonica from Indonesia (bulu types), have
one or very few productive tillers. Culm (stem) length is highly
variable, with the culms of domesticated deepwater rice being
able to grow to a length of 5 or 6 m .
Only three traits, culm length, panicle length and spikelet
shattering, were consistently recorded in six studies that
investigated QTLs for domestication related traits [49,57–62].
Secondary branching of panicles, spikelets per panicle and
heading date were recorded in ﬁve of these studies.
Thus, studies of the domestication syndrome in rice have
identiﬁed important QTL’s for domestication—but these QTL’s
are related to the parents of hybrid populations. After
domestication subsequent diversiﬁcation has resulted in
domestication related traits varying greatly from varietal group
to varietal group and within varietal groups. Further under-
standing of the domestication syndrome in rice will require
more crosses designed to dissect speciﬁc traits associated with
domestication in different varietal groups.
4.2. Single or multiple origins of Asian rice
The pendulum has swung in the last 30 years from the view
that rice was domesticated once [6,43,63] to being domesticated
two or more times [18,64–67].InBox 1 we list recent research
that appears to support single or dual/multiple origins of
domesticated rice. The list numerically favours dual/multiple
origins. Here we discuss this issue with reference to the criteria to
distinguish single from dual/multiple origins of crops elaborated
Box 1. Evidence supporting single and multiple events
leading to domesticated rice
1. The sh4 allele that is most important in reducing
shattering in rice, accounting for about 70% of the
loss in shattering. This loss of shattering results from
a functional base pair mutation that is the same in the
various groups of rice [2,4].
2. The red pericarp locus, rc, that affects change from
red pericarp, ubiquitous in wild rice, to white pericarp
common in all types of domesticated rice, does not
show segregation in crosses between japonica and
indica varieties. This recessive allele leading to white
pericarp is common in nearly all rice varieties (97.9%)
3. The bottleneck of domestication is strong in rice
[28,65,69]. While a severe genetic bottleneck does
not rule out multiple domestication one explanation
for this surprising result is that domestication of rice
was a geographically local event  that would
support a single domestication event.
Dual or multiple domestication
1. Crosses between indica and japonica rice result in
progeny that segregate for wild alleles at several loci
and wild characteristics reappear .
2. Indica, japonica, aromatic and aus varieties of culti-
vated rice tend to form monophyletic groups and
contain unique alleles suggesting they are derived
from unrelated wild populations [18,31].
3. Comparison of cytoplasmic diversity in wild and
cultivated rice suggests multiple lines of evolution
4. Genotyping and gene sequencing of cultivated and
wild rice has suggested that indica and japonica
accessions are related to different accessions of wild
5. Genomic divergence between indica and japonica
predates domestication and has been estimated to
have occurred 0.4–0.2 Mya [18,39,112].
6. Phylogeographic analysis of three genes in a broad
set of wild and cultivated rice accessions revealed
indica-like halotypes associated with wild rice from
South and Southeast Asia and japonica-like halop-
types associated with wild accessions from China
7. Haplotypes diversity analysis of genes reveals dis-
tinct nodal clusters of haplotypes associated primar-
ily with indica and japonica varieties [84,105].
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 399
by Zohary . He considered that in general domestication of
the same species twice is uncommon, at least in the Middle East.
He suggested three tests to discriminate single from multiple
origins under domestication :
Presence or absence of patterns indicative of founder effects
in the cultivated gene pools, compared to the amount of
variation in the wild progenitor.
Uniformity or lack of uniformity (within a crop) in genes
governing principal domestication traits.
4.2.1. Founder effects
In rice the domestication bottleneck has been studied based
on (a) nucleotide diversity data for 10 unlinked nuclear loci
; (b) haplotype diversity of one chloroplast and two nuclear
genes ; (c) genome-wide patterns of nucleotide poly-
morphism . These provide somewhat different pictures of
the rice domestication bottleneck. The gene-based studies
suggest a strong domestication bottleneck while the genome-
wide study suggested differences in patterns of diversity
between wild and cultivated rice as a result of both a genetic
bottleneck and selective sweeps after domestication. However,
it is clear a small fraction of the rich genetic diversity of wild
rice that must have existed across Asia 10,000 years ago, was
involved in domestication. This suggests rice is a product of a
single or very few introductions into cultivation.
4.2.2. Domestication traits
The most important domestication trait in rice is the loss of
shattering because that results in rice being dependent on
humans for survival—a domesticate. The main shattering allele
on chromosome 4 is sh4 (synonymous with the gene SHA1 in
reference ), as it accounts for about 70% of the phenotypic
variance for shattering . Zohary  states ‘... if in all
cultivars of a crop, a given domestication trait is found to be
governed by the same major gene (or same combination of
genes), this uniformity suggests a single origin’ (italics by
Zohary). The non-shattering allele of sh4 results from the same
base pair change in both indica and japonica cultivars from
different parts of Asia  conforming to that requirement.
Another domestication related trait, but not a trait that is
critical for domestication, is the change from red pericarp (in
wild rice) to white pericarp (in most cultivated rices). All white
pericarp indica (169 accessions) and japonica (186 accessions)
analysed are the result of the same 14 bp deletion [70,71].
Sweeney et al.  state that ‘the presence of this deletion in
97.9% of white-grained rice varieties found throughout the world
today suggests either that the gene was dispersed during the early
phases of domestication and is common by descent in modern
varieties or that very strong, positive selection for the allele led to
its introgression and maintenance in already established gene
pools’ (italics are ours). The ﬁrst reason in the previous sentence
seems the simplest and most logical explanation. Sweeney et al.
 support independent origin of indica and japonica and
suggest ‘‘early agriculturalists ... moved (the rc mutation)
around the Himalayan mountain range that is found between the
proposed centers of indica and japonica domestication ..., and,
having traversed this substantial geographic barrier, was rapidly
introgressed into all major subpopulations of rice despite an
emerging fertility barrier’’. It seems more reasonable that the rc
mutation, that is thought to have originated in a japonica type
haplotype , arose in the Yangtze river valley and was rapidly
introgressed into rice of various genetic backgrounds growing in
If there had been independent domestication of indica and
japonica rice we would expect the most important domestica-
tion traits to be the result of different mutations. As the above
examples do not reveal different mutations for the two main
groups of rice this suggests a single origin of rice followed by
4.2.3. Species diversity
Zohary  further suggests that if there are different wild
progenitors that could be domesticated it is possible to determine
how many of these wild species were cultivated and then
domesticated. While there is strong evidence to support some
differentiation of indica and japonica genomes long before rice
domestication [18,39] there is no information regarding whether
these proto-indica and proto-japonica genomes evolved in
geographical isolation. Several studies havesuggested that indica
cultivars are derived from O. nivara and japonica cultivars from
O. ruﬁpogon [67,72,73]. However, O. nivara and O. ruﬁpogon
are a genetic continuum and not ‘good’ taxonomic species. In
addition, both thesewild species can be sympatric and gene ﬂow
is possible between them. Research has shown that there is a lack
of clear geographic differences in the A-genome wild rice with
japonica-like and indica-like haplotypes based on the analysis of
ten unlinked loci in nearly complete linkage disequilibria
[74,75]. Thus, it is not possible to draw a ﬁrm conclusion on the
importance of ‘species’diversity in relation to the origin of rice.
A further critical consideration is that if rice was domesticated
once with rapid dispersal of the key domestication gene within an
area of diverse wild rice wewould not expect indica and japonica
groups of varieties have a different time span after domestication.
Indeed recent modelling of the bottleneck of domestication based
on genome-wide patterns of nucleotide polymorphism in rice has
suggested that equal timing of domestication has occurred for
indica and tropical japonica .
Based on the above comments rice appears to be the result of
a single domestication event despite many studies – using
present day germplasm – that point to dual or multiple
domestication of rice. The discrepancy can be explained by
post-domestication introgression, selection and diversiﬁcation
that are reﬂected in the current rice germplasm.
Below the scientiﬁc basis to support a hypothesis that can
resolve the seemingly contradictory results regarding how rice
was domesticated is presented.
4.3. The domestication of Asian rice
4.3.1. Diversity of wild and cultivated rice in China
There have been many studies of the diversity of populations
of A-genome wild rice, O. ruﬁpogon, in China [25,76–81].
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408400
Chinese wild rice populations are declining due to habitat
destruction and degradation, resulting in a high level of genetic
differentiation among populations . Wild rice in China
includes populations with diverse population structures from
perennial, essentially clonally, propagated populations to
mainly seed propagated populations . Wild rice at different
locations within a particular site has different characteristics
. The usual type of wild rice in China is mixed mating with
the ability to ratoon and also produce seeds [26,76]. The
ecotype of annual wild rice, O. nivara, originally described
from India has not been reported in China (Fig. 3b).
Wild rice in China consists of representatives of both indica-
like and japonica-like wild rice based on RFLP analysis .
Wild rice from Guangxi, where wild rice in China is most
diverse, had allozyme types characteristic of both indica and
japonica rice . The presence of japonica-like wild rice in
Guangxi is signiﬁcant since this is an area where most rice
production is indica type hence allozyme pattern in wild rice
might not be a simple reﬂection of introgression from sympatric
cultivated rice. Two groups of wild rice were found in China
using SSR markers that were differentiated based on
geography: one group consisted of populations of wild rice
from Hainan Island, and the other group from provinces to the
north . Chinese wild rice had a higher percentage of
polymorphic loci, more alleles, greater number of genotypes
and greater heterozygosity than wild rice from other regions in
a comparison of genetic diversity using RFLP variation of wild
rice from South Asia, Southeast Asia and China . China is
at the northern edge of A-genome wild rice diversity but it has a
very high level of genetic diversity.
The Yunnan province, China, is historically a major center of
traditional rice diversity , reﬂecting the effects of
ecogeographic variation and ethnic diversity . There is
no evidence to suggest that rice originated where it is
traditionally most diverse and genetic diversity of wild rice
in Yunnan is not as diverse as in other parts of China .
The allelic diversiﬁcation related to apiculus pigmentation
shows phenotypic divergence in cultivated rice but not wild rice.
Phenotypic divergence is due to the Clocus having multiple
alleles . The main group of haplotypes for the Clocus for
anthocyanin pigmentation at the apiculus is common in most
japonica and indica varieties. Two wild rice samples from China
included in the study were also in this group of haplotypes.
Cultivated Chinese rice also includes accessions belonging to a
second haplotype group similar to one other cultivated accession
from Laos, and both ‘nivara’ and ‘ruﬁpogon’ like wild rice from
Bangladesh, India and Thailand. This indicates a high level of
allelic variation in Chinese cultivated rice.
The presumed neutral nuclear pseudogene p-VATPase had a
haplotype diversity that reportedly reﬂected the geographic
pattern of the rice germplasm analysed . However,
cultivated rice accessions from China (9 accessions) and India
(13 accessions) had exactly the same array of haplotypes. The
other two genes analysed also had highly similar haplotype
diversity between Chinese and Indian cultivated rice .
As with wild rice, cultivated rice in China is highly diverse.
Chinese rice varieties span from indica to temperate japonica
from south to north and from indica to tropical japonica
cultivation in some parts of southern and southwestern
provinces where tropical japonica rice is grown by some
ethnic groups and at higher altitude.
4.3.2. An eco-genetic hypothesis for the domestication of
There is sufﬁcient archaeobotanical evidence to state that
rice was domesticated in the region of the Yangtze River valley
of China [85,86]. There is currently insufﬁcient archaebotanical
evidence to determine if rice was separately domesticated
elsewhere . The paleo-climate and paleo-ecology of China
is critical to understanding rice domestication. Prior to the
period when wild rice was cultivated and subsequently
domesticated, there were short and long cyclic periods of
warmer and cooler, wetter and dryer climates. Some of these
changes were a response to global climatic changes  others
reﬂected regional or local phenomena such as weather patterns
emanating from China’s Gobi desert region contributed
pulsations of arid and humid conditions . Vegetation shifts
across China over the last 10,000 years have been mapped and
reveal advancing and retreating patterns of warm temperate and
subtropical forests in the Yangtze River region . These
periodic changes in climatic conditions and advancing
retreating plant migrations probably contributed to the complex
genetic mosaic of locally adapted wild rice in China.
After the Younger Dryas (13,000–11,500 BP) climates
warmed and wild rice would have migrated north (Fig. 4).
However, beginning about 8000 BP the climate of China cooled
and monsoon rainfall declined. This mid-Holocene climate
change in China would have had an impact on wild rice
populations that likely declined and becoming locally extinct
and it has been suggested that this may have promoted a faster
rate of domestication  as other food sources such as nuts
were also in decline (Fig. 4).
Studies of vegetation history at Lake Daihan, west of
Beijing, have suggested that warm, humid weather patterns
between 3950–3500 BP and 1700–1350 BP . At these
times, after rice was domesticated, wild rice migrated as far as
the Yellow River and this is recorded in historic literature .
Wild rice stands were vast with wild rice being harvested in
various parts of China, with a record dated to 1100 BP for Hebei
(Hopei) province near Beijing that states ‘‘wild rice ripened in
an area of more than (13,333 ha), much to the beneﬁt of the
poor in local and neighbouring counties’’ . Thus, high seed
producing wild rice ecotypes existed and were harvested in
China long after rice itself was domesticated. However, since
domesticated rice was also being grown in the same area at this
time, opportunities for introgression would have been
abundant. Now wild rice in China is restricted to southern
and southwestern provinces (Fig. 5). Present wild rice
populations in China have introgressed with cultivated rice,
are rare, widely scattered and declining in both size and number
[77,81]. Therefore, contemporary Chinese wild rice germplasm
provides a poor reﬂection of what wild rice genetic diversity
might have been like in the past. In order to imagine wild rice in
the past in China, contemporary studies of adjacent areas of
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 401
Asia today, with lower human population density than China,
might provide a suitable comparison. The wild rice diversity on
the Vientiane Plain of Laos has been studied in depth [8,93,94].
Annual and perennial wild rice, including high seed producing
populations of O. ruﬁpogon, can be found scattered across the
Vientiane Plain .O. ruﬁpogon and O. nivara are usually
found at separate locations several kilometres apart and are
generally reproductively isolated by both distance and ﬂower-
ing time. Such an environment with high level of village-to-
village wild rice diversity would likely have existed along the
Yangtze River in the past.
The Yangtze River valley has a myriad of lakes where wild
rice must have grown in the past. The extent of these lakes and
the degree to which the water level rose and fell annually would
have determined the types of ecotypes of wild rice that evolved
there. These populations of wild rice were the material for the
domestication of rice. Post-Han dynasty records for wild rice in
Jiangxi province on the Yangtze River dated at 1300 BP state
‘‘In early spring ...(autumn?) ...wild rice ripened in an area of
1400 ha and perennial rices ripen in an area of (12,000 ha)’’
[92, p. 67]. This clearly indicates two distinctly recognisable
types of wild rice in the Yangtze valley in the past.
Therefore, the Yangtze River valley seems to have had
various ecotypes of O. ruﬁpogon that evolved in a similar
way to the complexity of wild rice that can currently be found
in the river plains of the Brahmaputra, Ganges, Irrawady,
Chao Pray and Mekong river valleys. Today both O.
ruﬁpogon and O. nivara grows in these river valleys of
South and Southeast Asia, but not in the valleys of the fast
ﬂowing Salween and Red rivers. When temperatures warmed
in China after the Younger Dryas the wild rice of the Yangtze
River valley might have been similar to that we ﬁnd in areas
with similar climates today, with annual and perennial
ecotypes of wild rice in different areas. This may explain the
high degree of residual genetic diversity found today in
Chinese wild rice (discussed above).
Hunters and gatherers would have been familiar with wild
rice as a source of nutritious seeds, and probably developed
methods similar to those used today to gather different ecotypes
of wild rice. Annual O. nivara is still harvested by tying plants
with leaves into bundles so that the grain falls into the centre of
the bundle not the ground, and the bundles are then harvested
(Fig. 1). Perennial O. ruﬁpogon, a taller plant, is harvested by
beating the panicles over a basket (Fig. 2). In deep water, where
O. ruﬁpogon grows, boats may have been used for harvesting,
as practiced by Native Americans to harvest Zizania (American
wild rice). Early rice harvesters and cultivators made boats ,
but we are not aware of wild rice being harvested by boat today
in Asia or Africa.
Fig. 5. How a single domestication event led to Asian rice evolving into highly
diverse varietal groups with deep genetic divisions. (A) Wild rice gene pool with
populations distributed across Asia (shaded light blue). The wavy line with
arrow represents the population from which domesticated rice ﬁrst evolved.
Other wavy lines represent populations with signiﬁcant variation that intro-
gressed into domesticated rice. (B) Finding of a non-shattering mutant. Dashed
line emanating from (B) represents the porous boundary to gene ﬂow between
the wild and domesticated rice gene pools. (C) Introgression at the wild/
domestication boundary resulted in diverging lines of domesticated rice that
in turn sometimes introgressed with one another, some lines becoming extinct,
leading to the present day genetic diversity of domesticated Asian rice indicated
by green. The more complex pattern of lines within domesticated rice compared
to wild rice represents movement due to transport by humans.
Fig. 4. Northern limit of past and present wild rice distribution in China.
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408402
A scenario can be envisaged where some groups went out
ﬁshing on lakes and harvested tall deepwater wild rice by one
method. While others, perhaps women and children, stayed
close to settlements where forests had been opened, harvested
short stature, annual wild rice from the more disturbed pools
and lake margins by a different method. The Mesolithic people
in the lower Yangtze manipulated and responded to their
environment with a level of sophistication . Between 8000
and 7000 BP there is evidence to suggest wetland management
that encouraged growth of wild grasses (including wild rice)
and reeds (Typha). There is also evidence that early forms
of domesticated rice were present in the same areas at about
7000 BP .
The Yangtze River valley (about 318N) is now the dividing
line between japonica and indica rice cultivation in China .
As the climate cooled between 7000 and 5000 BP some types of
rice would have adapted to cooler temperatures while others
adapted to warmer temperatures, as they were taken to southern
This scenario means indica-like and japonica-like wild rice
could have been gathered and subsequently cultivated in
parallel in the same region. The Yangtze River valley would
have provided natural zones of hybridisation and diversity that
were exploited during the process of rice domestication. The
key discovery of the mutation at the sh4 locus giving non-
shattering spikelets that resulted in domesticated rice could
then have quickly spread throughout this region where different
types of wild rice were gathered/cultivated. As discussed earlier
introgression may have been even more prevalent when rice
was ﬁrst domesticated than today because both inbreeding
reproductive system and barriers to hybridisation would not
have been well-developed. Gradual human migration with
domesticated rice would have been accompanied by introgres-
sion with wild rice in areas to which rice was spread leading to
distinct regional allelic proﬁles in rice. The successive cycles of
selection and introgression would have ﬁltered the best allelic
combinations that would have further spread. Our hypothesis
combines the elements of both models for domestication of rice
presented by Sang and Ge . Their models envisage either a
single origin of domesticated rice (the snowball model—for a
tropical crop!) or the multiple origins of domesticated rice
(combination model). While the important non-shattering allele
was found in one population (snowball model) the concurrent
and sympatric cultivation of different types of wild rice some
more japonica-like and some more indica-like would have
enabled this non-shattering allele to rapidly spread over a
restricted region while enabling the multiple types of rice in an
area to retain there other main characteristics. After the gradual
spread of the non-shattering allele resulting in the domestica-
tion of diverse wild rice ecotypes subsequent gene ﬂow would
have enabled, over time, other key domestication alleles to be
introgressed from one type of rice to another (combination
model). This would help to explain why the mutation that
resulted in the main non-shattering allele is common to indica
and japonica rice varieties today [2,4]. In addition, domestica-
tion of rice in a single region would account for the
domestication genetic bottleneck for rice [28,65,69].
Thus, deep genetic divisions that exist among rice cultivars
today can be explained by a single domestication event in a
geographic region of high wild rice diversity and subsequent
diversiﬁcation. The initial domesticated founder populations
that spread from a core region would have undergone cycles of
introgression (where wild rice occurred) and selection for genes
enabling rice to adapt to the new environments as it spread
(Fig. 5). Thus, all the evidence for dual and multiple
domestication presented in Box 1 can be explained by this
hypothesis. The interpretation that rice has been domesticated
more than once is a consequence of analysing present day
germplasm that has undergone several thousand years of rapid
evolution under human inﬂuence and impacted by the
components of the primary gene pool of rice.
Can this hypothesis be checked? It seems that accumulated
data on the diversity among varieties for key rice domestication
related genes—shattering and dormancy will be most important.
If a new mutation is found resulting in a new allele that causes
non-shattering via sh4, clearly single domestication hypothesis
for Asian rice will be untenable. Therefore, a search for such an
allele should be made. It is also likely that new techniques to
analyse increasing numbers of archaeological remains of rice at
the molecular level will help in the reconstruction of ancient wild,
cultivated and domesticated rice genetic diversity.
4.4. Evolution of rice in Africa
While there are perennial (O. longistaminata) and annual
(O. barthii) A-genome wild species in Africa, they do not have
the same close genetic relationship that is found between
perennial and annual wild rice of Asia . However, O.
longistaminata and O. barthii can commonly be found growing
in the same area. As discussed above, perennial O. long-
istaminata diverged early in the evolution of the A-genome.
Phylogenetic studies suggest the annual O. barthii evolved
from an ancestor of Asian wild rice not O. longistaminata
The domestication of rice in Africa occurred later than in
Asia but unequivocally prior to the introduction of Asian rice
[99,100]. The best evidence from archaeobotanical data for the
domestication of O. glaberrima comes from the site of Dia, in
the middle Niger Delta, and Mali . Abundant grains were
recovered from all levels at this site, the earliest occupation of
which was dated at between 2800 and 2500 BP. Based on
dimensions of grain and their lack of change in size over time
they were all presumed to be from domesticated plants. The
accelerator-based mass spectrometry C-14 dating of these rice
grains suggests that the time previously proposed for the
domestication of African rice of about 3500 BP  might be
close to the correct time.
Knowledge of the genetics of O. glaberrima is far less than
O. sativa although it is thought that the two rice species have
similar genetic architecture for many traits .O. glaberrima
differs from Asian rice by its more strictly annual habit, few
secondary panicle branches, short rounded ligule and red
pericarp. The domestication trait alleles in African rice have not
yet been compared to those of Asian rice.
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 403
O. glaberrima is the product of a double evolutionary
bottleneck. The ﬁrst associated with the divergence from Asian
Oryza, perhaps ancestors of O. barthii were introduced to
Africa from Asia. The second was due to domestication. This
helps to explain in part the lack of genetic diversity in O.
glaberrima compared to O. sativa . Although genetic
diversity of O. glaberrima is less than Asian rice, genetic
subgroups were detected in a large-scale SSR analysis of O.
glaberrima germplasm that was thought to reﬂect the
ecological specialization of O. glaberrima in deepwater, saline
affected and upland habitats . The domestication process
for African rice probably was similar to Asian rice, as the
methods used today to harvest wild rice are very similar in both
continents (cf. Figs. 1 and 2 with ﬁgures on page iv in reference
). However, the introduction of Asian rice in historic times to
Africa added a new dynamics to African rice culture. Today, a
proportion of African rice has introgressed genes from O. sativa
. An illustration of a simple evolutionary model for the
domestication of Asian and African rice is presented (Fig. 6).
5. Post-domestication—the waxy locus
Some cultures of Asia have selected glutinous forms of
cereals even for cereals not native to Asia such as maize .
Fig. 6. Scheme for the evolution of the cultivated rices. African rice has a restricted geographic distribution compared to Asian rice hence does not exhibit the
geographic variation that Asian rice.
Fig. 7. Evolution at the waxy locus (structure of waxy locus based on ): (a) point mutation leading to low amylose; (b) duplication event resulting in glutinous
rice; (c) crossover between glutinous and non-glutinous rice at the waxy locus between exons 1 and 2 resulting in progeny of glutinous rice without the point mutation
at the ﬁrst intron junction (c-1) and non-glutinous without the second exon duplication (c-2).
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408404
Wild rice does not have glutinous seeds. The mutation that
resulted in glutinous rice was selected after domestication. The
waxy locus is one of the most intensively studied regions of the
rice genome because it is associated with the taste and texture of
rice. This locus provides much information regarding how an
important mutation spreads in a crop gene pool. It also provides
information related to the impact of human selection at a
The alleles Wx
(normal high amylose rice) and Wx
amylose rice) differ by a single base pair mutation (G to T) at
the 50junction of the ﬁrst intron , and it was considered
that this mutation was required for glutinous rice .
However, at the same time six indica and ﬁve tropical japonica
varieties out of 353 glutinous varieties analysed did not have
this mutation . Using many tropical japonica varieties of
germplasm from the glutinous rice zone of Asia, a 23 bp
duplication in the 2nd exon of the waxy gene was found that is
characteristic of tropical glutinous rice  (Fig. 7). Based on
varieties screened to date, it is usually the 23 bp duplication that
is the cause of the glutinous trait in rice (, but for a rare
exception see reference ).
The evolution of glutinous rice seems to have occurred in
several stages (Fig. 7). The ﬁrst stage was the point mutation at
the 50junction of the ﬁrst intron that resulted in the waxy gene
being less effective and producing low amylose rice, the Wx
allele. This type of low amylose rice became common in
temperate japonica varieties. Subsequently, the 23 bp duplication
1146 bp down stream from the low amylose associated point
mutation occurred resulting in waxy rice with no amylose
production. Some time later a recombination event(s) must have
occurred between the point mutation and the 23 bp duplication
that can explain why some waxy rices (about 3%) do not have the
mutation at the 50junction of the ﬁrst intron but do have the 23 bp
duplication (S. Yamanaka, National Institute of Agrobiological
Sciences, Japan, 2007, personal communication).
The genomic region around the waxy locus has shed light on
the strong impact of selection on important traits. Clear evidence
of a selective sweep, reduced variation among nucleotides
neighboring a mutation as a result of strong selection was
reported for the waxy locus . The selective sweep showed
reduced nucleotide variation in a 250 kb region around this locus.
A comparison of the inferred selection coefﬁcient of this region
was made with gene regions in maize. While the selection
coefﬁcient was of the same order of magnitude in a similarly crop
diversifying gene region in maize (Y1 related to endosperm
colour), it was of a much higher order of magnitude than that for a
principal maize domestication related gene tb1 (teosinte
branched 1). The selective sweep at the waxy locus reveals
the strong impact of human selection on crop gene evolution. As
a result of the 10,000–8000 years that rice has been cultivated and
selected by humans, rice cultivars express genes adapted to the
needs of humans in ﬁnely organised arrays.
Results of genetic studies into the evolution of rice are
biased by the small fraction of the historical rice diversity that
now exists and is available for use from gene banks. Gene bank
germplasm represents a secondary genetic ‘footprint’ of past
Oryza variation in both time and space. Recent research
suggests that Australian A-genome germplasm is important for
understanding the evolution of Oryza A-genome diversity.
There has been insufﬁcient germplasm collection and analysis
of Australian Oryza genetic resources. Current day rice
germplasm has deep genetic divisions, not only indica and
japonica varieties but also other varietal groups such as aus and
aromatic varieties. While wild rice had some attributes
associated with indica and japonica rice prior to domestication
the genetic diversity in rice today largely reﬂects post-
domestication events. The deep genetic divisions in rice result
from selection within domesticated rice and continual
introgression from wild rice. The single mutation in
domesticated rice for the major non-shattering gene and the
severe genetic bottleneck associated with rice domestication
supports the view that Asian rice results from a single
domestication event, at least for the major varietal groups.
Initial domestication of rice would, however, have occurred in
an area with a high level of wild rice diversity. We hypothesise
that a rapid early spread of the main non-shattering gene of
domestication in rice, and possibly other domestication traits,
into wild rice representing different adaptive syndromes and
subsequent introgression as domesticated rice was spread to
new areas, can explain the deep divisions and diversity that we
see today in the global rice germplasm collection. A dual
domestication of indica and japonica rice in totally different
geographic regions is not congruent with current data on rice
domestication and key domestication related traits. African rice
evolved from a common ancestor with Asian rice that had
already diverged from the perennial African rice species O.
longistaminata. While there are parallels between the evolution
of Asian and African rice, African rice evolved from a more
restricted wild gene pool and its evolution in historic times has
been inﬂuenced by introduced Asian rice.
Post-domestication human inﬂuence on the rice genome is
well illustrated by studies of the waxy locus that affects taste
and texture of rice. Human selection has a major impact on
genomic regions where important crop divergence genes are
found and can result in useful mutations spreading widely in the
crop gene pool.
Despite so much interest and research on the evolution of
rice, it is surprising how much is still not known. Among areas
of priority research is further understanding of dormancy, that is
a complex trait consisting of kernal and hull components.
Another important area of rice evolution research concerns de-
domestication and how to prevent the occurrence and spread of
weedy rice. Insights into reversal of some aspects domestica-
tion exhibited by various types of weedy rice might be valuable
for understanding domestication itself. Thus, studies of all the
QTL associated with both shattering and dormancy should be a
priority. In addition, the imbalance in our understanding of
domestication in Asian rice compared to African rice needs
addressing given the array of useful genes for adaptation to
West African environments in O. glaberrima. The current
momentum to understanding rice evolution will continue as
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 405
new technologies are applied to study rice at all levels from
molecule to whole plant. These new data from rice research will
require synthesis with data from a broad range of disciplines to
obtain a more precise understanding of rice evolution.
In conclusion we believe the event that resulted in the origin
of the domesticated rice in Asia and Africa was simple, it was
subsequent processes of domestication that makes the story
The authors express their gratitude to Professor Song Ge,
Institute of Botany, Chinese Academy of Sciences, for
constructive discussions and review of an earlier version of
this paper. The authors also appreciate help from Dr. Dorian
Fuller, Institute of Archaeology, University College, London,
for bringing a number of references to their attention. The
advice and comments of several anonymous reviewers
including the review editor, Professor Jonathan Gressel, is
 S. Konishi, T. Izawa, S.-Y. Lin, K. Ebana, Y. Fukuta, T. Sasaki, et al., An
SNP caused loss of seed shattering during rice domestication, Science
312 (2006) 1392–1396.
 C. Li, A. Zhou, T. Sang, Rice domestication by reduced shattering,
Science 311 (2006) 1936–1939.
 H.-S. Ji, S.-H. Chu, W. Jiang, Y.-I. Cho, J.-H. Hahn, M.-Y. Eun, et al.,
Characterization and mapping of a shattering mutant in rice that corre-
sponds to a block of domestication genes, Genetics 173 (2006) 995–
 Z. Lin, M.E. Grifﬁth, X. Li, Z. Zhu, L. Tan, Y. Fu, et al., Origin of seed
shattering in rice (Oryza sativa L.), Planta 226 (2007) 11–20.
 J.R. Harlan, J.M.J. de Wet, Towards a rational classiﬁcation of cultivated
plants, Taxon 20 (1971) 509–517.
 H.I. Oka, Origin of Cultivated Rice, Elsevier, Amsterdam, 1988, pp. 1–
 L.J. Chen, D.S. Lee, Z.P. Song, H.S. Suh, B.R. Lu, Gene ﬂow from
cultivated rice (Oryza sativa) to its weedy and wild relatives, Ann. Bot.
93 (2004) 67–73.
 Y. Kuroda, Y.I. Sato, C. Bounphanousay, Y. Kono, K. Tanaka, Gene ﬂow
from cultivated rice (Oryza sativa L.) to wild Oryza species (O. ruﬁpogon
Griff. and O. nivara Sharma and Shastry) on the Vientiane plain of Laos,
Euphytica 142 (2005) 75–83.
 Y.E. Chu, H.I. Oka, Introgression across isolating barriers in wild and
cultivated Oryza species, Evolution 24 (1970) 135–144.
 M. Semon, R. Nielsen, M.P. Jones, S.R. McCouch, The population
structure of African cultivated rice Oryza glaberrima (Steud.): evidence
for elevated levels of linkage disequilibrium caused by admixturewith O.
sativa and ecological adaptation, Genetics 169 (2005) 1639–1647.
 B.R. Lu, M.E.B. Naredo, A.B. Juliano, M.T. Jackson, Taxonomic status
of Oryza glumaepatula Steud. III. Assessment of genomic afﬁnity among
AA genome species from the New World, Asia and Australia, Genet. Res.
Crop Evol. 45 (1998) 215–223.
 T. Ogawa, Genome research in genus Oryza, in: J.S. Nanda, S.D. Sharma
(Eds.), Monograph on Genus Oryza, Science Publishers, Inc., Plymouth,
UK, 2003, pp. 171–212.
 D.A. Vaughan, E. Balazs, J.S. Heslop-Harrison, From domestication to
super-domestication, Ann. Bot. 100 (2007) 893–901.
 R. Ikeda, Y. Sokei, I. Akintayo, Reliable multiplication of seeds for
NERICA varieties of rice, Genet. Resour. Crop Evol. 54 (2007) 1637–
 M.P. Jones, M. Dingkuhn, G.K. Aluko, M. Semon, Interspeciﬁc Oryza
sativa L. O. glaberrima Steud. progenies in upland rice improvement.
Euphytica 94 (1997) 237–246.
 C. Martin, A. Juliano, H.J. Newbury, B.-R. Lu, M.T. Jackson, B.V. Ford-
Lloyd, The use of RAPD markers to facilitate the identiﬁcation of Oryza
species within a germplasm collection, Genet. Res. Crop Evol. 44 (1997)
 Z.Y. Wang, G. Second, S.D. Tanksley, Polymorphism and phylogenic
relationships among species in the genus Oryza as determined by
analysis of nuclear RFLPs, Theor. Appl. Genet. 83 (1992) 565–581.
 Q. Zhu, S. Ge, Phylogenetic relationships among A-genome species of
the genus Oryza revealed by intron sequences of four nuclear genes, New
Phytol. 167 (2005) 249–267.
 K. Doi, M.N. Nonomura, A. Yoshimura, N. Iwata, D.A. Vaughan, RFLP
relationships of A-genome species in the genus Oryza, J. Fac. Agr.
Kyushu Univ. 45 (2000) 83–98.
 A.B. Juliano, M.E.B. Naredo, M.T. Jackson, Taxonomic status of Oryza
glumaepatula Steud. I. Comparative morphological studies of New
World diploids and Asian AA genome species, Genet. Resour. Crop
Evol. 45 (1998) 197–203.
 D.A. Vaughan, Wild Relatives of Rice: Genetic Resources Handbook,
IRRI, Los Banos, the Philippines, 1994, pp. 1–157.
 M. Akimoto, Y. Shimamoto, H. Morishima, Genetic differentiation in
Oryza glumaepatula and its phylogenetic relationships with other AA
genome species, Rice Genet. Newsl. 14 (1998) 37–39.
 A. Ghesquierre, Evolution of Oryza longistaminata, in Rice Genetics,
IRRI, Manila, Philippines, 1986, pp. 15–25.
 Y. Sano, H. Morishima, H.I. Oka, Intermediate perennial–annual popu-
lations of Oryza perennis found in Thailand and their evolutionary
signiﬁcance, Bot. Mag. Tokyo 93 (1980) 291–305.
 Z.P. Song, W. Zhu, J. Rong, X. Xu, J. Chen, B.R. Lu, Evidences of
introgression from cultivated rice to Oryza ruﬁpogon (Poaceae) popula-
tions based on SSR ﬁngerprinting: implications for wild rice differentia-
tion and conservation, Evol. Ecol. 20 (2006) 501–522.
 H.W. Cai, X.K. Wang, H. Morishima, Comparison of population genetic
structure of common wild rice (Oryza ruﬁpogon Griff.), as revealed by
analyses of quantitative traits, allozymes, and RFLPs, Heredity 92 (2004)
 P. Barbier, H. Morishima, A. Ishihama, Phylogenetic relationships of
annual and perennial wild rice: probing by direct sequence, Theor. Appl.
Genet. 81 (1991) 693–702.
 Q. Zhu, X. Zheng, J. Luo, B.S. Gaut, S. Ge, Multilocus analysis of
nucleotide variation of Oryza sativa and its wild relatives: severe bottle-
neck during domestication of rice, Mol. Biol. Evol. 24 (2007) 875–888.
 F. Bray, Joseph Needham Science and Civilization in China, vol. 6,
Biology and Biological Technology, Part II Agriculture, Cambridge
University Press, Cambridge, 1984, pp. 1–724.
 J.C. Glaszmann, Isozymes and classiﬁcation of Asian rice varieties,
Theor. Appl. Genet. 74 (1987) 21–30.
 A.J. Garris, T.H. Tai, J. Coburn, S. Kresovich, S.R. McCouch, Genetic
structure and diversity in Oryza sativa L., Genetics 169 (2005) 1631–
 O.F. Linares, From tidal swamp to inland valley: on the social organiza-
tion of wet rice cultivation among the Diola of Senegal, Africa 51 (1981)
 K. Ramiah, Rices of Madras: A Popular Handbook, The Superintendent,
Govt., Madras, India, 1937, pp. 1–249.
 F. Bray, The Rice Economies: Technology and Development in Asian
Societies, University of California Press, Berkeley, 1986, pp. 1–254.
 J. Gressel (Ed.), Crop Ferality and Volunteerism, Taylor & Francis, Boca
Raton, 2005, pp. 1–422.
 D.A. Vaughan, S. Ge, A. Kaga, N. Tomooka, Phylogeny and biogeo-
graphy of the genus Oryza, in: A. Hirai, T. Sasaki, Y. Sano, H. Hirano
(Eds.), Rice Biology in the Genomics Era, Biotechnology for Forestry
and Agriculture, vol. 62, Springer, 2008, pp. 219–234.
 V. Prasad, C.A.E. Stromberg, H. Alimohammadian, A. Sahni, Dinosaur
coprolites and the early evolution of grasses and grazers, Science 310
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408406
 Y.L. Guo, S. Ge, Molecular phylogeny of Oryzeae (Poaceae) based on
DNA sequences from chloroplast, mitochondrial, and nuclear genomes,
Am. J. Bot. 92 (2005) 1548–1558.
 J. Ma, J.L. Bennetzen, Rapid recent growth and divergence of rice
nuclear genomes, Proc. Natl. Acad. Sci. U.S.A. 101 (2004) 12404–
 F. Ren, B.R. Lu, S. Li, J. Huang, Y. Zhu, A comparative study of genetic
relationships among the AA-genome Oryza species using RAPD and
SSR markers, Theor. Appl. Genet. 108 (2003) 113–120.
 T. Ishii, T. Nakano, H. Maeda, O. Kamijima, Phylogenetic relationships
in AA genome species of rice as revealed by RAPD analysis, Genes
Genet. Syst. 71 (1996) 195–201.
 K.C. Park, N.H. Kim, Y.S. Cho, K.H. Kang, J.K. Lee, N.S. Kim, Genetic
variations of AA genome Oryza species measured by MITE-AFLP,
Theor. Appl. Genet. 107 (2003) 203–209.
 T.T. Chang, The origin, evolution, cultivation, dissemination and diver-
siﬁcation of Asian and African rices, Euphytica 25 (1976) 435–441.
 G.S. Khush, Origin, dispersal, cultivation and variation of rice, Plant
Mol. Biol. 35 (1997) 25–34.
 D.A. Vaughan, K. Kadowaki, A. Kaga, N. Tomooka, On the phylogeny
and biogeography of the genus Oryza, Breed. Sci. 55 (2005) 113–122.
 R.E. Huke, E.H. Huke, Rice, Then and Now, IRRI, Manila, Philippines,
1999, pp. 1–44.
 D.A. Vaughan, P.L. Sanchez, J. Ushiki, A. Kaga, N. Tomooka, Asian rice
and weedy rice—evolutionary perspectives, in: J. Gressel (Ed.), Crop
Ferality and Volunteerism, Taylor & Francis, Baco Raton, 2005, pp. 257–
 J. Lu, T. Tang, H. Tang, J. Huang, S. Shi, C.-I. Wu, The accumulation of
deleterious mutations in rice genomes: a hypothesis on the cost of
domestication, Trends Genet. 22 (2006) 126–131.
 H.W. Cai, H. Morishima, QTL clusters reﬂect character association in
wild and cultivated rice, Theor. Appl. Genet. 104 (2002) 1217–1228.
 X.-Y. Gu, Z.-X. Chen, M.E. Foley, Inheritance of seed dormancy in
weedy rice, Crop Sci. 43 (2003) 835–843.
 X.-Y. Gu, S.F. Kianian, M.E. Foley, Multiple loci and epistases control
genetic variation for seed dormancy in weedy rice (Oryza sativa),
Genetics 166 (2004) 1503–1516.
 X.-Y. Gu, S.F. Kianian, M.E. Foley, Phenotypic selection for dormancy
introduced a set of adaptive haplotypes from weedy into cultivated rice,
Genetics 171 (2005) 695–704.
 X.-Y. Gu, S.F. Kianian, G.A. Hareland, B.L. Hoffer, M.E. Foley, Genetic
analysis of adaptive syndromes interrelated with seed dormancy in
weedy rice (Oryza sativa), Theor. Appl. Genet. 110 (2005) 1108–1118.
 X.-Y. Gu, S.F. Kianian, M.E. Foley, Isolation of three dormancy QTLs as
Mendelian factors in rice, Heredity 96 (2006) 93–99.
 X.-Y. Gu, S.F. Kianian, M.E. Foley, Dormancy genes from weedy rice
respond divergently to seed development environments, Genetics 172
 S.L. Dillon, F.M. Shapter, R.J. Henry, G. Corderio, L. Izquierdo, L.S.
Lee, Domestication to crop improvement: Genetic resources for Sorghum
and Saccharum (Andropogoneae), Ann. Bot. 100 (2007) 975–989.
 H.D. Catling, D.W. Puckridge, D. HilleRisLambers, The environments of
Asian deepwater rice, in: 1987 International Deepwater Rice Workshop,
IRRI, Manila, Philippines, 1988, pp. 11–34.
 C. Bres-Patry, M. Lorieux, G. Cle
´ment, M. Bangratz, A. Ghesquie
Heredity and genetic mapping of domestication-related traits in a
temperate japonica weedy rice, Theor. Appl. Genet. 102 (2001) 118–126.
 S.J. Lee, C.S. Oh, J.P. Suh, S.R. McCouch, S.N. Ahn, Identiﬁcation of
QTLs for domestication-related and agronomic traits in an Oryza sativa x
O. ruﬁpogon BC
population, Plant Breed. 124 (2005) 209–219.
 C. Li, A. Zhou, T. Sang, Genetic analysis of rice domestication syndrome
with the wild annual species, Oryza nivara, New Phytol. 170 (2006) 185–
 K. Onishi, Y. Horiuchi, N. Ishigoh-Oka, K. Takagi, N. Ichikawa, M.
Maruoka, et al., A QTL cluster for plant architecture and its ecological
signiﬁcance in Asian wild rice, Breed. Sci. 57 (2007) 7–16.
 L.Z. Xiong, K.D. Liu, X.K. Dai, C.G. Xu, Q. Zhang, Identiﬁcation of
genetic factors controlling domestication-related traits of rice using an
F2 population of a cross between Oryza sativa and O. ruﬁpogon, Theor.
Appl. Genet. 98 (1999) 243–251.
 B.R. Lu, K.L. Zheng, H.R. Qian, J.Y. Zhuang, Genetic differentiation of
wild relatives of rice as assessed by RFLP analyses, Theor. Appl. Genet.
106 (2002) 101–106.
 D.Q. Fuller, Agricultural origins and frontiers in South Asia: a working
synthesis, J. World Prehist. 20 (2006) 1–86.
 J.P. Londo, Y.C. Chiang, K.H. Hung, T.Y. Cheng, B. Schaal, Phylogeo-
graphy of Asian wild rice reveals multiple independent domestications of
cultivated rice, Oryza sativa, Proc. Nat. Acad. Sci. U.S.A. 103 (2006)
 G. Second, Origin of the genic diversity of cultivated rice (Oryza spp.):
study of the polymorphism scored at 40 isozyme loci, Jpn. J. Genet. 57
 J.-H. Xu, C. Cheng, S. Tsuchimoto, H. Ohtsubo, E. Ohtsubo, Phyloge-
netic analysis of Oryza ruﬁpogon strains and their relations to Oryza
sativa strains by insertion polymorphism of rice SINEs, Genes Genet.
Syst. 82 (2007) 217–229.
 D. Zohary, Monophyletic vs. polyphyletic origin of the crops on which
agriculture was founded in the Near East, Genet. Resour. Crop Evol. 46
 A.L. Caicedo, S.H. Williamson, R.D. Hernandez, A. Boyko, A. Fiedel-
Alon, T.L. York, et al., Genome-wide patterns of nucleotide polymorph-
ism in domesticated rice, PLoS Genet. 3 (2007) 1745–1756.
 M.T. Sweeney, M.J. Thomson, B.E. Pfeil, S. McCouch, Caught red-
handed: Rc encodes a basic helix-loop-helix protein conditioning red
pericarp in rice, Plant Cell 18 (2006) 283–294.
 M.T. Sweeney, M.J. Thompson, Y.G. Cho, Y.J. Park, S.H. Williamson,
C.D. Bustamante, et al., Global dissemination of a single mutation
conferring white pericarp in rice, PLoS Genet. 3 (2007) 1418–
 S. Yamanaka, I. Nakamura, K.N. Watanabe, Y.I. Sato, Identiﬁcation of
SNPs in the waxy gene among glutinous rice cultivars and their evolu-
tionary signiﬁcance during the domestication process, Theor. Appl.
Genet. 108 (2004) 1200–1204.
 C. Cheng, R. Motohashi, S. Tsuchimoto, Y. Fukuta, H. Ohtsubo, E.
Ohtsubo, Polyphyletic origin of cultivated rice: based on the intersper-
sion pattern of SINEs, Mol. Biol. Evol. 20 (2003) 67–75.
 T. Tang, J. Lu, J. Huang, J. He, S.R. McCouch, Y. Shen, et al., Genomic
variation in rice: genesis of highly polymorphic linkage blocks during
domestication, PLoS Genet. 2 (2006) 1824–1833.
 T. Tang, S. Shi, Molecular population genetics of rice domestication, J.
Integr. Plant Biol. 49 (2007) 769–775.
 L.Z. Gao, S.G.D. Hong, Allozyme variation and population genetic
structure of common wild rice Oryza ruﬁpogon Griff. in China, Theor.
Appl. Genet. 101 (2000) 494–502.
 L.Z. Gao, B.A. Schaal, C.H. Zhang, J.Z. Jia, Y.S. Dong, Assessment of
population genetic structure in common wild rice Oryza ruﬁpogon Griff.
using microsatellite and allozyme markers, Theor. Appl. Genet. 106
 Z.P. Song, X. Xu, B. Wang, J.K. Chen, B.R. Lu, Genetic diversity in the
northernmost Oryza ruﬁpogon populations estimated by SSR markers,
Theor. Appl. Genet. 107 (2003) 1492–1499.
 Z.P. Song, B. Li, J.K. Chen, B.R. Lu, Genetic diversity and conservation
of common wild rice (Oryza ruﬁpogon) in China, Plant Species Biol. 20
 Z.W. Xie, Y.Q. Lu, S. Ge, D.Y. Hong, F.Z. Li, Clonality in wild rice
(Oryza ruﬁpogon Poaceae) and its implications for conservation manage-
ment, Am. J. Bot. 88 (2001) 1058–1064.
 H. Zhou, Z. Xie, S. Ge, Microsatellite analysis of genetic diversity and
population genetic structure of a wild rice (Oryza ruﬁpogon Griff.) in
China, Theor. Appl. Genet. 107 (2003) 322–339.
 C.Q. Sun, X.K. Wang, Z.C. Li, A. Yoshimura, N. Iwata, Comparison of
the genetic diversity of common wild rice (Oryza ruﬁpogon Griff.) and
cultivated rice (O. sativa L.) using RFLP markers, Theor. Appl. Genet.
102 (2001) 157–162.
 M. Nakagahra, Geographic distribution of esterase genotypes of rice in
Asia, Rice Genet. Newsl. 1 (1984) 118–120.
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408 407
 K. Saitoh, K. Onishi, I. Mikami, K. Thidar, Y. Sano, Allelic diversiﬁca-
tion at the C(OsC1) locus of wild and cultivated rice: Nucleotide changes
associated with phenotypes, Genetics 168 (2004) 997–1007.
 D.Q. Fuller, E. Harvey, L. Qin, Presumed domestication? Evidence for
wild rice cultivation and domestication in the ﬁfth millennium BC of the
Lower Yangtze region, Antiquity 81 (2007) 316–331.
 Y.F. Zheng, G.P. Sun, X.G. Chen, Characteristics of the short rachillae of
rice from archaeological sites dating to 7000 years ago, Chin. Sci. Bull.
52 (2007) 1654–1660.
 L. Wang, M. Sarnthein, H. Erlenkeuser, P.M. Grootes, J.O. Grimalt, C.
Pelejero, G. Linck, Holocene variations in Asian monsoon moisture: a
bidecadal sediment record from the South China Sea, Geophys. Res. Lett.
26 (1999) 2889–2892.
 R.O. Whyte, The origins of the Chinese environment, in: D.N. Keightley
(Ed.), The Origins of Chinese Civilization, University of California
Press, 1983, pp. 3–19.
 G. Ren, H.J. Beug, Mapping Holocene pollen data and vegetation of
China, Quart. Sci. Rev. 21 (2002) 1395–1422.
 D.Q. Fuller, Contrasting patterns in crop domestication and domestica-
tion rates: recent archaeobotanical insights from the Old World, Ann.
Bot. 100 (2007) 903–924.
 J. Xiao, Q. Xu, T. Nakamura, X. Yang, W. Liang, Y. Inouchi, Holocene
vegetation in the Daihai Lake region of north-central China: a direct
indication of the Asian monsoon climatic history, Quart. Sci. Rev. 23
 P.-T. Ho, The Cradle of the East, The Chinese University of Hong Kong
and University of Chicago Press, 1975, pp. 1–444.
 Y. Kuroda, H. Urairong, Y.I. Sato, Differential heterosis in a natural
population of Asian wild rice (Oryza ruﬁpogon Griff.) due to reproductive
strategy and edge effect, Genet. Resour. Crop Evol. 52 (2005) 151–160.
 Y. Kuroda, Y.I. Sato, C. Bounphanousay, Y. Kono, K. Tanaka, Genetic
structure of three Oryza AA genome species (O. ruﬁpogon O. nivara and
O. sativa) as assessed by SSR analysis on the Vientiane Plain of Laos,
Conserv. Genet. 8 (2007) 149–158.
 L. Jiang, L. Liu, The discovery of an 8000-year-old dugout canoe at
Kuahuqiao in the lower Yangzi River, China. Antiquity 79 Project
Gallery (2005) http://antiquity.ac.uk/projgall/liu/index.html.
 Y. Zong, Z. Chen, J.B. Innes, C. Chen, Z. Wang, H. Wang, Fire and ﬂood
management of coastal swamp enabled ﬁrst rice paddy cultivation in east
China, Nature 449 (2007) (2007) 459–463.
 J.-H. Shen, Rice breeding in China, in Rice improvement in China and
other Asian countries, IRRI-CAAS, Manila, Philippines, 1980, pp. 9–30.
 T. Sang, S. Ge, The puzzle of rice domestication, J. Integr. Plant Biol. 49
 M. Sweeney, S. McCouch, The complex history of the rice domestica-
tion, Ann. Bot. 100 (2007) 951–957.
 D.A. Vaughan, S. Miyazaki, K. Miyashita, The rice gene pool and human
migrations, in: D. Werner (Ed.), Biological Resources and Migration,
Springer, Berlin, 2004, pp. 1–13.
 S.S. Murray, Searching for the origins of African rice domestication.
Antiquity 78 (2004) at http://antiquity.ac.uk/projgall/murray/index.html.
 R. Porteres, African cereals: Eleusine, Fonio, Black Fonio, Teff, Bra-
chiaria, Paspalum, Pennisetum, and African rice, in: J.R. Harlan, J.M. de
Wet, A.B.L. Stemler (Eds.), Origins of Plant Domestication, Mouton
Publishers, The Hague, 1976, pp. 409–452.
 S. Sakamoto, Glutinous-endosperm starch food culture speciﬁc to East-
ern and Southeastern Asia, in: R. Ellen, K. Kikuchi (Eds.), Redeﬁning
Nature. Ecology, Culture and Domestication, Oxford, UK, 1996, pp.
 H.Y. Hirano, M. Eiguchi, Y. Sano, A single base change altered the
regulation of the Waxy gene at the posttranscriptional level during the
domestication of rice, Mol. Biol. Evol. 15 (1998) 978–987.
 K.M. Olsen, M.D. Purugganan, Molecular evidence on the origin and
evolution of glutinous rice, Genetics 162 (2002) 941–950.
 S. Wanchana, T. Toojinda, S. Tragoonrung, A. Vanavichit, Duplicated
coding sequence in the waxy allele of tropical glutinous rice (Oryza
sativa L.), Plant Sci. 165 (2003) 1193–1199.
 T. Inukai, A. Sako, H.Y. Hirano, Y. Sano, Analysis of intragenic
recombination at wx in rice: Correlation between the molecular and
genetic maps within the locus, Genome 43 (2000) 589–596.
 Y. Hori, R. Fujimoto, Y. Sato, T. Nishio, A novel wx mutation caused
by insertion of a retrotransposon-like sequence in a glutinous
cultivar of rice (Oryza sativa), Theor. Appl. Genet. 115 (2007)
 K.M. Olsen, A.L. Caicedo, N. Polato, A. McClung, S. McCouch, M.D.
Purugganan, Selection under domestication: Evidence for a sweep in the
rice waxy genomic region, Genetics 173 (2006) 975–983.
 S. Kawakami, K. Ebana, T. Nishikawa, Y.I. Sato, D.A. Vaughan, K.
Kadowaki, Genetic variation in the chloroplast genome suggests multiple
domestication of cultivated Asian rice (Oryza sativa L.), Genome 50
 S. Rakshit, A. Rakshit, H. Matsumura, Y. Takahashi, Y. Hasegawa, A. Ito,
et al., Large-scale DNA polymorphism study of Oryza sativa and O.
ruﬁpogon reveals the origin and divergence of Asian rice, Theor. Appl.
Genet. 114 (2007) 731–743.
 C. Vitte, T. Ishii, F. Lamy, D. Brar, O. Panaud, Genomic paleontology
provides evidence for two distinct origins of Asian rice (Oryza sativa L.),
Mol. Gen. Genomics 272 (2004) 504–511.
D.A. Vaughan et al. / Plant Science 174 (2008) 394–408408