# Paddy rice cultivation

> Rice grown in level, bunded, flooded fields

- Date: c. 6,200 BCE
- Era: Prehistory
- Domain: [Agriculture & Livestock](https://lastcodex.org/en/domains/agriculture/)
- Rebuild stage: II (Settlement)
- Difficulty: 2/5
- URL: https://lastcodex.org/en/inventions/paddy-rice-cultivation/

A paddy is a level field closed by low earth banks, called bunds, and kept under a few centimetres of water while rice grows. A channel brings the water in through a cut in the bund, the soil is churned into mud so that the water does not drain away, and the field is emptied before harvest. The oldest probable rice fields, found with their ditches at Hanjing in eastern China, are 8,400 to 8,000 years old.

![Paddy rice cultivation · Rice grown in level, bunded, flooded fields](https://lastcodex.org/_astro/plate.CYkyU9Gh_ZqYJmM.webp)

## Legend

1. Feed channel
2. Inlet cut in the bund
3. Earth bund
4. Standing water a few centimetres deep
5. Puddled mud
6. Compacted hardpan

## What it is

![A block of hillside with three curved rice fields stepping down it, each holding a shallow sheet of water with rows of young rice standing in it; water from a small channel at the top, then from each field, spills through a notch in the bank into the field below](https://lastcodex.org/_astro/v3.8ydKx-Nn_LyVsX.webp)
_Water comes from a channel or the field above through a cut in the bund and passes on to the field below._

Rice descends from a semi-aquatic wild plant, and the International Rice Research Institute (IRRI) notes that it suffers quickly from lack of water. A paddy is a field levelled flat and closed on every side by a low bank of packed earth, the bund, so that it can hold a sheet of water a few centimetres deep for most of the season. Water arrives from a channel or from the field above through a cut in the bund and leaves the same way into the field below or a drain.

Wild rice grew in marshes long before anyone farmed it. With bunds, levelled ground, puddled soil and channels, the paddy lets farmers decide how wet the rice is at each stage of its growth. Paddies cover river plains, deltas and, as [terraces](/en/inventions/terrace-farming/), whole mountainsides.

## How it works

The standing water keeps the roots supplied through dry spells and holds back the weeds, which are smothered by the flood while the transplanted rice stands above it. IRRI advises about 3 cm of water after transplanting, rising to 5–10 cm as the plants grow, and draining the field 7–10 days before harvest.

For the water to stay, the field must be level and nearly watertight. Farmers flood the cultivated field and look at the surface: wherever soil shows above the water there is a high spot to smooth, and the water level is then lowered step by step to show the next ones. The bunds must be well packed, with cracks and rat holes plastered with mud, and high enough not to overflow in heavy rain; IRRI asks for at least 20 cm. Permanent bunds also serve as paths. FAO describes them as 60 to 90 cm high when built, settling to 40 to 50 cm over several months.

Clay soils suit paddies best because water passes through them slowly. Before transplanting, farmers work the soaked soil until it turns into a soft mud, which is called puddling. It destroys the structure of the soil, which matters little for rice, and IRRI explains that puddled lowland fields develop a compacted hardpan below the mud, which reduces the water lost downwards. On sandy soils puddling helps little, and FAO notes that rice there loses much more water unless the water table is high.

On a gentle slope the fields form a chain. Water enters the highest one, fills it, spills through a notch into the next and so on down, which suits clay soils that lose little water. IRRI gives an average of 1,432 litres of water to produce a kilogram of irrigated lowland rice, and estimates that irrigated rice takes 34–43% of the world's irrigation water.

## How to do it

A small plot of two by two metres on clay or clay loam is enough to try the method during a warm season. Raise the seedlings first, for a few weeks, in a tray of wet soil.

Dig out the topsoil inside the plot and heap it into a bank all round. Tread and pat the bank firm, wet it and plaster its inner face with mud until it stands at least 20 cm above the floor of the plot. Leave a notch in the bank at the side nearest your water supply and close it with a clod. Fill the plot with water and leave it to soak overnight. The next day, work the wet soil with a spade and your feet until it turns into smooth mud with no lumps, then smooth it with a board. Let the water settle and look for mud showing above the surface; push it into the low places until the water lies evenly everywhere.

![A block of rice-field soil cut open: a perforated tube driven into it, only its top sticking a little above the cracked, dry surface, and water standing at the bottom of the tube well below that surface; rice clumps with their roots on either side](https://lastcodex.org/_astro/v1.WaL8meZC_2rUuDT.webp)
_The tube shows where the water stands below the surface once the field looks dry._

Bring the water down to about 3 cm. Plant the seedlings in small bunches at regular spacing, pushing the roots a few centimetres into the mud. As they grow, raise the water to 5–10 cm. Top it up through the notch whenever it falls, and mend any leak in the bank with mud at once.

To save water, you can try IRRI's alternate wetting and drying once the plants are established, one to two weeks after transplanting. Take a tube about 30 cm long and 10 to 15 cm across, pierce it all over with holes, and drive it into the soil so that 15 cm stands above the surface. Scoop the soil out from inside. Let the water drop until it stands 15 cm below the surface inside the tube, then flood the field again to 5 cm. Keep the field flooded from a week before flowering to a week after. Drain the plot 7 to 10 days before harvest.

## History and evolution

At Hanjing, on the Huai River plain in Jiangsu, a team from the National Museum of China, the Nanjing Museum and the Sihong County Museum found three field-like structures and 18 small ditches next to a moated village of the Shunshanji culture. A three-dimensional model of the ground showed water flowing towards them, and an ancient stream channel nearby may have served both to feed and to drain them. Phytoliths of domesticated rice, microscopic silica bodies from the plant, lie inside them. Zhenwei Qiu and his colleagues describe them as probable rice fields or modified wild habitats of rice, from the first phase of the site, which falls within 8,400 to 8,000 years ago. In January 2025, the archaeological institute of Hunan announced two rice paddies more than 8,000 years old at Lijiagang, in the middle Yangtze region.

At Kuahuqiao, near Hangzhou, about 7,700 years ago, Yongqiang Zong and his colleagues found in the pollen and charcoal record that people burned the alder scrub of a coastal marsh to grow rice, and they argue that the flooding was probably controlled with bunds. At Tianluoshan, phytolith studies point to rice grown in wet conditions close to those of wild rice, 6,800 to 6,500 years ago. At Chengtoushan, on the middle Yangtze, a field of about 4400 BCE, 2.7 m wide and over 20 m long, ran between two earth banks.

![A reconstruction of a Neolithic field system: six small dug-out fields of different sizes separated by low earth ridges and linked by narrow channels to a deeper round pit full of water, with a pottery jar at the edge of the pit](https://lastcodex.org/_astro/v2.p_grukeF_1jcfzs.webp)
_Small fields joined to a water pit, like those excavated at Caoxieshan and Chuodun._

The oldest excavated field systems with channels and reservoirs come from east of Lake Tai, at Caoxieshan and then Chuodun. Dorian Fuller and Ling Qin date them to the late Majiabang period, 4300–4000 BCE; Rubi Wu and her colleagues give 5,950 to 5,700 years ago for Caoxieshan. They are small hollows dug into the subsoil, at Chuodun from about one to fifteen square metres, joined by channels to deeper pits that held water in reserve. Fuller and Qin suggest that water was lifted from the pits by bucket and let in and out of the fields through the channels. At Chenghu, in the Songze period (4000–3300 BCE), some fields approach 100 square metres and lie in more regular rows. At Maoshan, in the Liangzhu period, large paddies framed by regular channels and ridges were built about 4,700 years ago and farmed until about 4,000 years ago.

According to Fuller and Qin, rice was present at many sites of the Yellow River basin by or after 3000 BCE, with older finds at Jiahu and Yuezhuang and a fourth-millennium one at Nanjiaokou, and it reached central Thailand and Cambodia by about 2000 BCE. They also relate this spread to a rise in atmospheric methane after 3000 BCE, for which the climatologist William Ruddiman and his colleagues see rice paddies as a likely source. Today, alternate wetting and drying, promoted by IRRI, lets fields dry for days at a time without lowering the yield, to save irrigation water.

## Key facts

- Oldest probable rice fields: Hanjing, Jiangsu, China, 8,400 to 8,000 years ago
- Oldest field systems with channels and reservoirs: Caoxieshan and Chuodun, near Lake Tai, 4300–4000 BCE
- Water depth after transplanting (IRRI): About 3 cm, rising to 5–10 cm as the rice grows
- Water used (IRRI): 1,432 litres per kilogram of irrigated lowland rice on average
- Irrigated rice's share of the world's irrigation water (IRRI): 34–43%

## Requires

- [Stone Tools](https://lastcodex.org/en/inventions/stone-tools/): The first cutting edge (3.3 million years ago)
- [Seed selection and saving](https://lastcodex.org/en/inventions/seed-saving/): Choosing plants, keeping the next crop alive (c. 8,500 BCE)
- [Irrigation](https://lastcodex.org/en/inventions/irrigation/): Leading water to the crops (c. 6,000 BCE)

## Sources

- [Excavation at Hanjing site yields evidence of early rice cultivation in the Huai River more than 8000 years ago (accepted manuscript)](https://discovery.ucl.ac.uk/id/eprint/10146677/1/Manuscript.pdf), Zhenwei Qiu, Lina Zhuang, Huiyun Rao, Liugen Lin, Yijie Zhuang, Science China Earth Sciences / UCL Discovery
- [Water management and labour in the origins and dispersal of Asian rice (World Archaeology 41, 2009, pp. 88–111)](https://doi.org/10.1080/00438240802668321), Dorian Q. Fuller, Ling Qin, World Archaeology
- [Rice stable carbon and nitrogen isotope values as potential indicators of watering and manuring regimes (2024)](https://doi.org/10.3389/fearc.2024.1488754), Rubi Wu et al., Frontiers in Environmental Archaeology
- [Fire and flood management of coastal swamp enabled first rice paddy cultivation in east China (Nature 449, 2007)](https://doi.org/10.1038/nature06135), Yongqiang Zong et al., Nature
- [Ecology and hydrology of early rice farming: geoarchaeological and palaeo-ecological evidence from the Late Holocene paddy field site at Maoshan, the Lower Yangtze (2019)](https://doi.org/10.1007/s12520-018-0639-1), Yuxi Jin et al., Archaeological and Anthropological Sciences
- [Paddy-field remains dating back 8,000 years discovered in China (Xinhua, 21 January 2025)](https://www.jsmsj.org.cn/JiangsuArtists/English/Culture/2025/0123/10437.html), Xinhua, republished by the Jiangsu Provincial Art Museum
- [Water management](http://www.knowledgebank.irri.org/step-by-step-production/growth/water-management), IRRI Rice Knowledge Bank, International Rice Research Institute
- [Saving Water with Alternate Wetting Drying (AWD)](http://www.knowledgebank.irri.org/training/fact-sheets/water-management/saving-water-alternate-wetting-drying-awd), IRRI Rice Knowledge Bank, International Rice Research Institute
- [Irrigation Water Management Training Manual No. 5, Irrigation methods, chapter 2, Basin irrigation](https://www.fao.org/4/s8684e/s8684e03.htm), Food and Agriculture Organization of the United Nations
- [Japanese encephalitis, fact sheet (6 August 2024)](https://www.who.int/news-room/fact-sheets/detail/japanese-encephalitis), World Health Organization

Fact-checked on 2026-10-18.

---

The Last Codex · https://lastcodex.org/en/inventions/paddy-rice-cultivation/ · © The Last Codex. Texts, plates and figures: CC BY-NC-ND 4.0.