All crops
Drag to turn the globe
10,000 yr ago6,000 yr ago4,000 yr ago2,000 yr ago1000 CE17001950Today

Crop Nº 04Poaceae (grasses)

Rice

Oryza sativa and Oryza glaberrima

Two wild marsh grasses, tamed on two continents. In China the taming took thousands of years.

Rice is two crops. Asian rice, Oryza sativa, and African rice, Oryza glaberrima, are different species, tamed separately from different wild grasses.

In China the taming took thousands of years, and the plant went on changing as it traveled: into indica and japonica, sticky and fragrant, tall and dwarf. It is now the staple of more people than any other food.

2
species, tamed separately in Asia and in Africa
820 M
metric tons grown in 2024, in the husk
89%
of it grown in Asia
172 M
hectares harvested in 2024
  • vrīhiSanskrit, rice; first found in the Atharvaveda
  • *pajayProto-Austronesian, rice in the field
  • *beRasProto-Austronesian, husked rice
  • óryzaGreek, the source of the English word
  • maloDyula (Mande), uncooked rice

Rice is the work of the first farmers of the lower Yangtze and of West Africa, and of the farmers who have kept its varieties since. The genebank at Los Baños holds more than 130,000 of them.

14 chapters · 158 sources

Numbers open the source. Their color shows how settled the claim is:Established Supported Contested

Before farming

A grass of the marshes

The wild plant sprawls, sheds its seed and hides a red grain in a barbed husk.

Wild rice, collected near Saigon in 1886. The panicle is open and every grain carries a long awn. Specimen collected by B. Balansa; photographed by Naturalis Biodiversity Center, Herbarium specimen L.1277089, Oryza rufipogon Griff., collected 31 October 1886. Naturalis Biodiversity Center, Leiden. CC0 1.0. None.

Rice began as a wild grass of wet ground, Oryza rufipogon, which grows across the tropics and subtropics of Asia. It comes in two forms: a tall perennial, and a short annual that some botanists count as a separate species, Oryza nivara.

It is a hard plant to harvest. It sprawls along the ground. Its grains fall as they ripen, each one armed with a long barbed awn, and the grain inside is red.

Long before anyone farmed it, wild rice had split into separate lineages. The two that lead to today's indica and japonica rice parted hundreds of thousands of years ago.

24, in 12 pairs
Chromosomes
about 50% · about 4%
Seed set by crossing: the perennial form, the annual
23
Species in the genus Oryza
12,000 to 8,000 years ago

Rice before farming

People gathered rice, potted with its husks and perhaps brewed with it, long before it changed.

The oldest traces of rice in human hands are microscopic: silica bodies from rice plants, in two caves in Jiangxi, 12,000 to 9,000 years old. They are usually read as wild rice, gathered.

By about 9,400 years ago people at Shangshan, in Zhejiang, were mixing rice husks into their potting clay. Residues in their jars have been read as rice beer. Whether the rice was already changing under their hands is argued. The evidence is the shape of cells from the leaf, and critics say that cannot show whether the ears still shattered.

The oldest pieces of rice ear come from Huxi, nearby, and are 9,000 to 8,400 years old. Six in ten are of the wild kind that sheds its grain. Fewer than one in ten held on. Rice was one food among many: the same jars held acorn, Job's tears and lily.

61% · 30% · 9%
Rice ear fragments at Huxi: wild, in between, tamed
seventh millennium BCE
A drink of rice, honey and fruit at Jiahu
9,000 to 6,000 years ago

A slow taming

Ears that held their grain crept from a few in a hundred to most, over thousands of years.

A wild ear sheds its grain. People who harvest and resow favor the rare ear that holds on, and the change can be counted in the broken stalks left behind.

At Kuahuqiao, about 7,700 years ago, people burned off swamp scrub and kept a wet grassland of the kind rice likes, until the sea drowned it. At Tianluoshan, between 6,900 and 6,600 years ago, the share of ears that held their grain rose from 27% to 39%, and rice went from 8% to 24% of the plant remains. Acorns and water chestnuts still filled the storage pits.

When the change was complete is argued: about 6,000 years ago by one reckoning. And one site spoils the neat story. At Baligang, far inland, most ears already held their grain by 6700 to 6300 BCE.

27% → 39%
Ears that held their grain at Tianluoshan, 300 years apart
8% → 24%
Rice among the plant remains there
6 of 15
Storage pits at Hemudu filled with acorns and water chestnuts
Over some three thousand years

The plant, rebuilt

It held its grain, stood up, lost its barbs, grew a heavier head and turned white.

Domestication changed rice in a handful of ways, and most are now traced to single genes. The grain stopped falling. One changed amino acid in sh4 weakened the layer of cells that lets a ripe grain drop, and in japonica a second change, in qSH1, added to it. The plant stood up: wild rice sprawls, and a broken PROG1 gene makes it grow erect. The awns went, through at least four genes.

The head changed too. It closed up, grew more branches, and came to carry about three times as many grains.

And the grain turned white. Wild rice is red. Nearly all white rice carries the same small deletion, 14 letters long, in a gene called Rc.

◂ The models beside this text are live. Drag to turn them.

about 58 → 175–203
Spikelets on a panicle: wild Asian rice, cultivated
330 of 337
White varieties explained by one deletion in Rc
182
Cultivars from 17 countries with the same changes in PROG1
7,000 to 4,300 years ago

Fields of water

From dug plots two meters across to a walled city behind dams.

A jade cong of the Liangzhu culture, about 2400 BCE. Liangzhu culture, Ritual object (cong), about 2400 BCE. The Metropolitan Museum of Art, New York (accession 2004.52). CC0 1.0. Resized.

The first rice fields were small and wet. At Tianluoshan, excavators describe marsh cleared with fire and spades about 7,000 years ago. The oldest field systems usually cited, at Chuodun and Caoxieshan, date to 4200–3800 BCE. Their plots were less than two meters across. The work at first was draining the ground, not watering it.

The fields grew. In its late phase the field at Maoshan covered about 5.5 hectares, in squares of about 1,000 square meters. Twenty kilometers away stood Liangzhu, a walled city of about 300 hectares, served by earthen dams. Only in its time did rice become the staple. In the middle of the city lay a store of burned rice: 10 to 15 metric tons, by one archaeologist's estimate.

Liangzhu ended about 4,300 years ago, in a short spell of unusually heavy rain.

under 2 m across
The earliest field plots
about 5.5 ha
The late field at Maoshan
10–15 t, 100 t or 200 t
Burned rice at Liangzhu, in different reports
3700 BCE to 800 BCE

North to Korea and Japan

Rice entered millet country as a guest, and needed a new kind of plant to go farther.

Setting out rice seedlings, by Hokusai, about 1795. Katsushika Hokusai, Spring in the Rice Fields, about 1795. The Metropolitan Museum of Art, New York (accession JP2546). CC0 1.0. Cropped to the print, leaving out the museum's color bar, and resized.

North of the Yangtze was millet country. Rice was grown on the Wei River by about 5,700 years ago, on the evidence of plant silica. Charred grain says later. Either way it stayed a minor crop there.

About 4,200 years ago the climate cooled, and by one genome study japonica then split into a temperate rice and a tropical one. To ripen in northern summers rice also had to lose most of its sense of day length.

Rice farming is firmly shown in Korea from about 1300 BCE, with many paddy fields. It reached northern Kyushu, in Japan, around the 9th century BCE, some 500 years after Korea.

nearly half
Rice among the crop grains in a pit at Liangchengzhen, about 2000 BCE
53° N to 40° S
Latitudes at which rice is grown today
2800 BCE to 1000 BCE

South by land and sea

To Taiwan, the Philippines and Thailand, and no farther into the Pacific.

Planting out rice in the Padang Highlands of Sumatra, about 1900–1920. Christiaan Benjamin Nieuwenhuis, Indische boeren aan het werk op een rijstveld, op de achtergrond de bergen van Sumatra, about 1900–1920. Rijksmuseum, Amsterdam (RP-F-F01149-DA). Public domain. Cropped to the photograph, leaving out the album page and its handwritten caption, and resized.

Rice left the Yangtze slowly, and by one account only once the craft of the wet field had been worked out.

By 2800–2200 BCE farmers at Nanguanli, on the west coast of Taiwan, grew rice and millet. Taiwan is the homeland of the Austronesian languages, and their ancestor already had three words for rice: one for the plant in the field, one for the husked grain, one for the cooked.

A husk in a potsherd at Andarayan puts rice on Luzon by about 2000 BCE, on the strength of one date. On the mainland, dry-field rice moved through the southern hills, and the first tamed rice in Thailand, at Khok Phanom Di, dates to 2000–1500 BCE. It was japonica.

In the islands rice counted for less, and the sailors who went on into the Pacific left it behind for taro and tree crops.

*pajay · *beRas · *Semay
Rice in Proto-Austronesian: in the field, husked, cooked
3,935–3,378 years ago
Rice husk in a potsherd at Andarayan, Luzon
About 2000 BCE

A second beginning?

India had rice of its own. How much of it was tamed there is still argued.

India had rice long before Chinese rice arrived. Charred grains at Lahuradewa, on the Ganges plain, date to about 6400 BCE. Whether they were farmed, tended or simply gathered is argued. Rice of clearly tamed type appears in the north only around 1800 BCE, at Mahagara.

In between, about 2000 BCE, japonica rice came in. One model holds that it crossed with the local half-tamed rice, passed on the genes that hold the grain on the ear, and so made indica.

The genomes half agree. Indica, japonica and a third Indian rice, aus, descend from three different wild lineages. One team finds japonica's genes in the other two, and so one true domestication. Another finds none at the genes that matter, and so three.

◂ The family tree is beside this text. Point at a form to trace its line.

about 4,200 years
Between the first rice and tamed rice in northern India
1 of 1,789
Indica samples with japonica's version of all nine domestication genes
about 17% · about 15%
Japonica ancestry in indica and in aus, by one estimate
By 800 to 400 BCE

Africa's own rice

A different wild grass, tamed in West Africa by its own farmers.

African rice from northern Nigeria, 1929. The label calls it cultivated swamp rice. Specimen collected by H. C. Sampson; photographed by Naturalis Biodiversity Center, Herbarium specimen WAG.1242684, Oryza glaberrima Steud., collected 19 December 1929. Naturalis Biodiversity Center, Leiden (Wageningen herbarium, a duplicate from Kew). CC0 1.0. None.

Rice was tamed a second time, in West Africa, from a different wild grass. Oryza barthii is an annual of seasonal floodwater, with awns up to 19 centimeters long and grain that falls so readily that people who gather it cut the heads unripe.

One genome study finds the wild stands shrinking for more than 10,000 years as the Sahara dried, and suggests that scarcity pushed people to sow. In 1962 Roland Portères placed the cradle in the inland delta of the Niger, in Mali. Some geneticists agree. Others see several beginnings across West Africa.

The oldest grains come from Dia, in that delta, and date to 800–400 BCE. The plant changed as Asian rice had, by different mutations: the grain held, the stem stood up, the awns went. The grain stayed red.

mostly 8–16 cm
The awn of Oryza barthii
about 47 → about 152
Spikelets on a panicle: wild African rice, tamed
800–400 BCE
Oldest dated African rice, at Dia
500 BCE to 1475 CE

The long road west

A rarity in Persia and Rome, a field crop in Muslim Spain, and a cargo to Madagascar.

Rice went west as a rarity. Tablets from Persepolis show it grown along the Persian royal road by about 500 BCE. Alexander's companions saw it in India, standing in water, and heard of it from Bactria to Syria. Theophrastus, about 300 BCE, described a grain grown mostly in water and boiled to a mash.

Rome knew it as an import. Ships carried rice up the Red Sea, and 196 charred grains lay in an army hospital on the Rhine.

It stayed marginal until the early Islamic centuries. Farming manuals from Muslim Spain describe sowing and transplanting it, though only one grain has ever been dug up there. In 1475 the Duke of Milan promised twelve sacks of seed to Ferrara.

By sea it went farther. Rice was the main crop of the first farming villages on the Comoros and Madagascar, in the 8th to 10th centuries.

196
Rice grains in the Roman army hospital at Neuss
70–100%
Rice among food crops on the Comoros and Madagascar, 650–1200 CE
miriziš
Old Persian word for rice, borrowed into Elamite
1012 CE

A faster rice

In 1012 a Song emperor sent to Fujian for a rice that ripened early.

Four men tread a chain pump to lift water into the rice. A Chinese handscroll painted before 1353. Unidentified Chinese artist, Rice Culture, or Sowing and Reaping (detail), before 1353. The Metropolitan Museum of Art, New York (accession 2005.277). CC0 1.0. None. This is the museum's own detail photograph.

In 1012 drought struck the rice lands of the Yangtze and the Huai. The Song emperor Zhenzong had heard of a rice from Champa, in what is now central Vietnam, that ripened early and stood drought. He sent to Fujian for 30,000 bushels of seed and had it handed out with instructions.

Nobody knows when the rice had reached Fujian, and historians disagree about how much the order itself changed. The new rice yielded less at first, and farmers selected better kinds. One sample tested in our own time proved to be aus, the early rice of eastern India.

Quick rice and terraced hills eased the food supply of Song China, and drew still more people to the rice-growing south. A Song official, Lou Shou, painted the farming year in 21 scenes. The set was copied for centuries, and the Kangxi emperor had a new one made in 1696.

30,000 bushels
Seed bought in Fujian in 1012
21
Scenes of rice farming in Lou Shou's Pictures of Tilling and Weaving
1690s to 1860

Across the Atlantic

Rice made Carolina rich. Who knew how to grow it is the argument.

Rice on the Ogeechee, near Savannah, Georgia, drawn for Harper's Weekly in January 1867. Alfred R. Waud, for Harper's Weekly, Rice culture on the Ogeechee, near Savannah, Georgia, 5 January 1867. Library of Congress, Prints and Photographs Division (LC-DIG-ds-07365). Public domain. Cropped to the page and resized.

Carolina's planters were trying rice by the 1690s. By 1770 Carolina and Georgia shipped more than 80 million pounds of it a year. Enslaved people grew it. Before about 1715 most were brought from the West Indies; later, from Africa.

Since the 1970s historians have argued that they brought more than labor: that people from the rice coasts of West Africa knew the crop, and their owners learned it from them. Critics answer with shipping records. Before 1751 about a fifth of the captives landed in Carolina and Georgia came from Upper Guinea, and the Caribbean took seven times as many people from rice country and exported no rice. The reply is that the records are thinnest for the years that matter.

The rice itself was Asian: Carolina Gold is a tropical japonica. African rice crossed the ocean too. Maroon farmers in Suriname still grow a little, and their oral history says it came from Africa hidden in a woman's hair.

0.4 → 80.8 million lb
Rice shipped from Carolina and Georgia, 1700 and 1770
21.9% → 58.2%
Captives from Upper Guinea among arrivals, before 1751 and 1751–75
4 of 136
Maroon rice varieties that are African rice
1962 to 1976

Miracle rice

A dwarf crossed with a giant in 1962 changed the rice of half of Asia.

The old rices of the tropics were tall and leafy. Fed with fertilizer, they fell over.

In 1962, at a new institute in the Philippines, Peter Jennings crossed Peta, a tall Indonesian rice, with Dee-geo-woo-gen, a dwarf from Taiwan, or from China; the accounts differ. A quarter of the grandchildren were short, so a single gene was at work. It proved to be a broken gene for making the growth hormone gibberellin.

Plant 3 of row 288 became IR8, named in November 1966. In trials that year it gave more than nine metric tons a hectare, where the Philippine average was about one. A Manila newspaper called it miracle rice. Its grain was chalky and it caught diseases, and hardier varieties followed.

China took another road. In 1970 Li Bihu found a male-sterile wild rice on Hainan, and hybrid rice was released in 1976.

9.4 t/ha
IR8 in trials at Los Baños, dry season of 1966
about 30% → about 70%
Asia's rice land under modern varieties, 1970 and 1990
50% → about 28%
Hybrid rice's share of China's rice land, 1994 and recently
Today

Rice today

Grown on 144 million farms, most of them smaller than a hectare.

Weeding the young rice while two men swing water into the field. From the imperial Pictures of Tilling and Weaving, 1696. Designed by Jiao Bingzhen for the Kangxi emperor, Rice cultivation: four men transplanting rice seedlings (?) in a paddy field, while behind them two men irrigate the field. Woodcut, 1696, 1696. Wellcome Collection, London (571282i). Public domain. Resized.

Rice is the staple of more people than any other food. In 2024 the world grew 820 million metric tons of it, in the husk, on 172 million hectares. That is nearly four times the harvest of 1961, from half as much land again. Nine-tenths is grown in Asia, and India and China grow half of it between them. Little of it travels: about 7% is exported. It comes from some 144 million farms, most of them smaller than a hectare.

Flooded fields have costs. They give off about 8% of the methane that people cause, and rice takes up far more arsenic from wet soil than other grains do.

In West Africa, Asian rice has largely replaced the African species, though Jola and Mende farmers still grow it for ritual. From 1992 breeders crossed the two. Their NERICA varieties were first released in 2000. In Uganda one study found they cut poverty, and another that more than half of the first growers gave them up within two years.

820 M
Metric tons grown in 2024
1.87 → 4.76 t/ha
World average yield, 1961 and 2024
132,627
Samples in the rice genebank at Los Baños