One scorching summer day in 1991, having spent hours surveying the biodiversity of sacred groves in southern West Bengal, India, I approached Raghu Murmu’s hut to rest. Raghu, a young man of the Santal tribe, sat me under the shade of a huge mango tree while his daughter fetched me cold water and sweets made from rice. As I was relishing these, I noticed that Raghu’s pregnant wife was drinking a reddish liquid. Raghu explained that it was the starch drained from cooked Bhutmuri rice—meaning “ghost’s head” rice, perhaps because of its dark hull. It “restores blood in women who become deficient in blood during pregnancy and after childbirth,” he said. I gathered that this starch is believed to cure peripartum anemia in women. Another rice variety, Paramai-sal, meaning “longevity rice,” promotes healthy growth in children, Raghu added.

As I would subsequently establish, Bhutmuri is one of several varieties of indigenous rice in South Asia that are rich in iron, and it also contains certain B vitamins. And Paramaisal rice has high levels of antioxidants, micronutrients and labile starch, which can be converted rapidly to energy. At the time, however, such uncommon rice varieties, with their evocative names and folk medicinal uses, were new to me. When I returned home to Kolkata, I conducted a literature survey on the genetic diversity of Indian rice and realized that I had been lucky to encounter Raghu. Farmers like him, who grow indigenous rice and appreciate its value, are as endangered as the varieties themselves.

In the years since, I have become familiar with a cornucopia of native rice varieties (also called landraces) that possess astonishingly useful and diverse properties. Some can withstand flood, drought, salinity or pest attacks; others are enriched in valuable vitamins or minerals; and yet others are endowed with an enticing color, taste or aroma that has given them special roles in religious ceremonies. Collecting, regenerating and sharing with farmers these exceedingly rare but valuable varieties has become my life’s mission.

Bringing back forgotten rice landraces requires the sowing, tending and harvesting of more than 1,000 varieties every year. Scenes from Basudha depict an indigenous farmer transplanting baby plants into a flooded field (top) and another working (bottom) on the farm. Credit: Zoë Savitz

Lost Treasure

Asian cultivated rice (Oryza sativa) resulted from centuries of selection and breeding of wild ancestral species—a process that Charles Darwin called “artificial selection”—by early Neolithic humans. Archaeological and genetic evidence suggests that the indica subspecies of Asian rice (almost all cultivated rice from the Indian subcontinent belongs to this group) was grown about 7,000 to 9,000 years ago in the foothills of the eastern Himalayas. Over the ensuing millennia of domestication and cultivation, traditional farmers created a treasure trove of landraces that were perfectly adapted to diverse soils, topographies and microclimates and suited to specific cultural, nutritional or medicinal needs.

According to pioneering rice scientist R. H. Richharia, more than 140,000 landraces were grown in India’s fields until the 1970s. If we exclude synonyms (that is, the same variety referred to by different names in different locales), this figure boils down to around 110,000 distinct varieties. As I learned from my literature survey, however, the genetic diversity of Indian rice has declined steeply since the advent of the Green Revolution.

In the late 1960s the International Rice Research Institute (IRRI) provided the Indian government with a few high-yielding varieties (HYVs) of rice, which provide substantial quantities of grain when supplied with ample water, fertilizer and pesticides. In concert with international development agencies, the IRRI urged the replacement of indigenous varieties across all types of fields with these imported strains. Heavily promoted and sometimes forced onto farmers’ fields, the new rice types rapidly displaced the landraces.

In the late 1970s and early 1980s IRRI researchers listed 5,556 landraces in West Bengal and collected 3,500 of these for its gene bank. In 1994, finding no documentation of surviving varieties in the state, I began my own, lone survey. Finally completed in 2006, it revealed that 90 percent of the documented varieties had vanished from farmers’ fields. In fact, it is likely that no more than 6,000 rice landraces exist in fields across India. Similarly, the Bangladesh Rice Research Institute documented the names of 12,479 varieties between 1979 and 1981, but my analysis of a recent study indicates that no more than 720 landraces are still cultivated in the entire country.

When I got an inkling of this staggering loss of biodiversity in the subcontinent, it shocked me as a biologist and as a concerned citizen. I wondered why agricultural institutions were unconcerned about the genetic erosion of the most important cereal of the region. After all, the dire consequences of the loss of genetic diversity of a key crop should have been evident from Ireland’s Great Famine of 1845–1849.

Most potatoes grown in Ireland were of a single variety, the Irish Lumper, which had no inherent resistance to Phytophthora infestans, the microorganism that causes potato blight. In 1846 three quarters of the harvest was lost to infection, resulting in a scarcity of seed potatoes in subsequent years and major demographic effects: up to 1.5 million people died from starvation and disease over the course of the famine, and in more than a decade of hunger and deprivation about 1.3 million people emigrated from Ireland to North America and Australia. The unforgettable lesson for agriculturists is that the absence of multiple varieties of a crop can make that plant vulnerable to pest or disease infestations: monocultures are disastrous for long-term food security. In the wake of the Green Revolution, insects such as the rice hispa and the brown planthopper, which had never before posed a significant problem, devastated rice crops in several Asian countries.

Vast expanses of monocultures provide banquets for certain pests. Farmers may try to eliminate them with generous applications of pesticides—which end up killing the natural enemies of those pests. The net effect is to enhance the diversity and abundance of pests, thus driving the pesticide mill wheel. The genetic uniformity of crop species—in particular the Green Revolution varieties, selected for the single trait of high yields—also means the plants lack endowments that would enable them to withstand vagaries of the weather such as insufficient or too late rain, seasonal floods or storm surges that inundate coastal farms with seawater. Their fragility makes a poor farmer who might not have the money to, say, buy a pump to irrigate his or her fields more vulnerable to environmental fluctuations.

The loss of landraces further entails the withering of a knowledge system associated with their cultivation. For example, traditional farmers can distinguish varieties by observing the flowering time; the color of the basal leaf sheath; the angle of the flag leaf; the length of the panicle; and the size, color and shape of the grain [see graphic below]. Using these and other characteristics, they eliminate all atypical or “off-type” plants to maintain the genetic purity of the landrace. Nowadays, however, the vast majority of South Asian farmers rely on an external seed supply, which obviates the need to conserve the purity of homegrown seeds. When a local variety is no longer available, the knowledge related to its agronomic and cultural uses fades from the community’s memory. Millennia-old strategies for using biodiversity to control pests and diseases have been supplanted by advice from pesticide dealers—to the detriment of soil and water quality, biodiversity and human health.

The Green Revolution and, more broadly, the modernization of agriculture have also had severe social and economic effects. Rising costs of inputs such as seeds, fertilizer, pesticides and fuel for irrigation pumps require farmers to borrow money, often from private money lenders. Debt, coupled with falling prices for the harvested crops, has contributed to distress sales of small farms and an epidemic of farmer suicides in India. In contrast, over decades of working with tribal farmers who are still growing local rice and millet varieties on their marginal farms, I have encountered not a single case of farm-related suicide.

In 1996, with 152 landraces in my collection, I approached the West Bengal State Directorate of Agriculture’s Rice Research Station, where all heirloom rice germplasm is supposed to be conserved. Not only did it refuse to accept and maintain the seeds I had collected, but the director chastised me for pursuing the “unscientific and retrogressive” goal of reviving the forgotten landraces. To insist on growing them would mean “going back to the caveman’s age” and condemning farmers to low productivity and lifelong poverty, he said. When I argued that none of the HYVs can survive on dryland farms without irrigation, on deep-water farms or on coastal saline farms, he assured me that modern transgenics would soon come up with the best varieties for those marginal farms, so I should leave the matter with the experts in agricultural science.

Living Seeds

Trained as an ecologist specializing in ecosystem structures and functions, I was working with the eastern regional office of World Wide Fund for Nature-India. At that time, it and other conservation organizations typically sought to safeguard large, charismatic animals such as the tiger, but because cultivated crops are not “wildlife,” there was no focus on their protection. Research institutions were also uninterested because the conservation of folk crop varieties would receive no funding support.

The only option left to me was to go it alone. I resigned from my job in 1996 and settled in a village in West Bengal to set up a folk rice seed bank and exchange center for farmers. In 1997 I named it Vrihi, Sanskrit for “broadcast rice.” In the early years I used my savings and considerable support from Navdanya, a New Delhi–based nongovernmental organization, to collect rare seeds from different corners of the country and distribute them for free to farmers in need. Since 2000, however, donations from friends and supporters have constituted the bulk of our funding.

In 1999, while in northern Bengal to survey biodiversity for the state’s forest department, I took the opportunity to explore the region’s fields. One day, after six hours of travel by bus and on foot to a remote village named Lataguri, I collected a critically endangered rice variety named Agni-sal. (I define a critically endangered variety as one that is being grown on only one farm.) The grain was fiery red in color—hence the name Agni, meaning “fire”—and its stem was strong enough to withstand storms. The next season I gave the seeds to a farmer who was looking for a rice that would flourish on his highland farm, which was swept by strong winds. He returned the following year with a broad smile of gratitude because of the great harvest from this rice, despite a cyclone that had devastated all the neighboring farms. The year after that, however, an officer from the district’s agriculture department persuaded him to replace Agni-sal with an HYV. As a result, Agni-sal was lost from our accession. I rushed to Lataguri to procure another sample from the original donor farmer, only to learn that he had passed away the year before and that his son had abandoned that rice. Agni-sal thus, to my knowledge, went extinct from the world.

Credit: Rebecca Konte

Another incident at about this time persuaded me that I needed to do more than collect and distribute seeds. Traditional lowland farmers in India used to grow two types of flood-tolerant rice. One can grow taller and taller in tandem with rising water levels. This underwater “stem elongation” property, governed by the genes SNORKEL 1 and SNORKEL 2, located on chromosome 12, is seen in traditional varieties such as Lakshmi dighal, Jabrah, Pantara and Rani kajal. A second type of flood-tolerant landrace can withstand prolonged submergence in floodwater. One of the genes governing submergence tolerance is SUB1, found in several Bengal landraces.

In June 1999 a southern district of West Bengal experienced a flash flood. All rice crops perished. At the time, my accession had no varieties that could tolerate submergence, but I knew that the IRRI and the National Bureau of Plant Genetic Resources in New Delhi possessed several dozen. I wrote to both institutions, requesting that they send me 10 to 20 grams of these seeds to save the distressed farmers. I received no acknowledgment from either of the gene banks. If an educated person, writing in a European language on letterhead showing his academic degrees and affiliations, does not merit any response from the national and international gene banks, one can imagine how likely it is that a poor farmer from Kenya or Bangladesh might receive seed samples from them. To my knowledge, no farmer in any country has ever received any seeds from these lofty ex situ, or off-site, gene banks—even though their accessions were built on contributions from traditional farmers.

In contrast, the gene banks do make their accessions available to seed companies for hybridization programs and patenting. An estimate by the International Food Policy Research Institute indicates that by 1996 about three quarters of U.S. rice fields had been sown with material descendant from the IRRI collection. And in 1997 the U.S. Patent and Trademark Office granted the broadest ever patent on an indigenous rice, for a hybrid strain of basmati whose parents originated in South Asia and were accessed from the IRRI collection, to Texas-based company RiceTec. The IRRI, which supposedly holds its accession in trust for the world’s farmers, itself applied in 2014 for an international patent on a yield-boosting rice gene called SPIKE discovered in the Indonesian landrace Daringan. (The governing body of the International Treaty on Plant Genetic Resources for Food and Agriculture has reviewed the legality of this controversial application but has yet to announce its decision.)

Not only are ex situ seed banks physically and socially distant from farmers, but also their seeds are handicapped by long isolation. Rice seeds are dried and preserved at –20 degrees Celsius, which keeps them viable for up to 35 years. Frozen in time, they are separated from the constantly evolving life-forms in the outer world. When grown out after 35 years, they will have lost any inherent resistance to specific pathogens, which will meanwhile have evolved into newer strains. In contrast, farmers’ in situ seed banks are necessarily low budget, so they must sow all the seeds every year—otherwise most of the rice would fail to germinate. Thanks to this imperative, the seeds conserved on farms continue to coevolve with diverse pathogens and pests.

After a series of such experiences and observations, I decided to set up a conservation farm of my own to maintain a small population of each landrace so that it would survive even if abandoned by most farmers. I used my savings from a postdoctoral fellowship at the University of California, Berkeley, to found Basudha farm in 2001. Vrihi is now South Asia’s largest open-access rice gene bank, and its 1,420 varieties are grown every year on Basudha farm in a tribal village in southern Odisha. Of the varieties in our accession, 182 are now extinct from India’s fields.

With less than 0.7 hectare at our disposal, we have to grow 64 individual plants of each variety on only four square meters of land. (The minimum number of plants required to maintain all the genetic endowments of a given landrace is about 50.) Because we cannot adhere to the internationally recommended isolation distance of at least 110 meters on every side of each landrace, preventing cross-pollination between neighboring varieties is a challenge. I managed to overcome this constraint by planting the different varieties so that each is surrounded by others with different flowering dates. Furthermore, we eliminate the off-type plants in each population at different life stages by observing 56 different characteristics, as per Bioversity International guidelines. After this step, all the seeds harvested are assumed to be 100 percent genetically pure, barring some undetected mutations.

On Basudha farm, all the rice landraces are grown in accordance with the agroecological principle of “zero external input”—no agrochemicals, no groundwater extraction, no fossil fuels. Nutrient supply comes from leaf and straw mulch, legume cover crops (whose roots are rich in nitrogen-fixing microbes), composted greens and animal manure, biochar and soil microbes. We control pests by growing “weed” grasses and shrubs that provide habitats for predators such as spiders, ants and reptiles, as well as parasites. Another strategy is to maintain puddles of water as breeding habitats for aquatic insects and frogs, which also prey on crop pests. And we occasionally use herbal pest repellents such as tobacco, garlic and tulsi (Ocimum sanctum; also known as holy basil). Crop diseases are never a problem on Basudha: varietal and species diversity has repeatedly been documented as the best strategy for protection against crop pathogens.

We store some of the harvested seeds in earthen pots, which protect them from insects and rodents while allowing them to “breathe,” for the next year’s sowing. The rest we distribute among farmers, in exchange for a handful of seeds of other folk varieties, which we cultivate and donate to farmers. This system is a conscious attempt to revive the ancient practice of seed exchange in all farming communities, which had once helped all crop varieties to spread across continents.

My co-workers and I have helped establish more than 20 other seed banks in different parts of India, so that local farmers can access the varieties they need without having to travel to Vrihi. We also promote seed-exchange networks among farmers. These banks and networks have benefited more than 7,800 farmers in five Indian states. Further, we document the characters and properties of each variety and register the landraces in the name of farmers to preclude any biopiracy patents on them. By such means, we seek to restore to farmers sovereignty over seeds—essential to their long-term financial and nutritional security.

Debal Deb and his long-term associate Debdulal Bhattacharya examine, record (top)
and discuss (bottom) the detailed characteristics of rice grains from the harvest.


On precarious farms experiencing drought or seasonal floods, traditional landraces are the only reliable means of providing food security to poor farmers. After 22 years of growing folk rice varieties, I am confident that landraces such as Kelas, Rangi, Gadaba, Kaya and Velchi will provide greater yields than any of the modern HYVs in drought conditions. Lakshmi dighal, Rani kajal and Jabra can elongate their stems as floodwaters rise, keeping their seed-bearing panicles above water up to four meters deep. Matla, Getu, Talmugur and Kallurundai can grow on saline soil and survive seawater incursion. These landraces are stable germ lines with a suite of genes conferring broad adaptive plasticity.

Moreover, given optimal soil conditions in rain-fed farms, a considerable number of folk rice varieties such as Bahurupi, Bourani, Kerala sundari and Nagra can outyield modern HYVs. A set of exceedingly rare varieties with relatively high yields includes double- and triple-kernel rice; these may have resulted from selections of rare mutations in the structural genes of the rice flower. Basudha seems to be the last repository of one such triple-kernel rice landrace, Sateen.

Several landraces also possess resistance to pests and pathogens. Kalo nunia, Kalanamak, Kartik-sal and Tulsi manjari are blast-resistant. Bishnubhog and Rani kajal are resistant to bacterial blight. Kataribhog is moderately resistant to tungro virus. Gour-Nitai, Jashua and Shatia seem to resist caseworm attack, and stem-borer attack on Khudi khasa, Loha gorah, Malabati, Sada Dhepa and Sindur mukhi varieties is seldom observed. Such seeds, distributed from Vrihi, have reduced crop losses from pest and disease attacks in thousands of farm fields over the past 25 or so years.

Modern rice breeding is largely focused on enhancing grain yield, but numerous folk rice varieties contain various micronutrients that are absent from modern cultivars. Our recent studies identified at least 80 folk varieties that contain more than 20 milligrams of iron per kilogram of rice, with the highest levels recorded for Harin kajli, Dudhé bolta and Jhuli rice, which range from 131 to 140 milligrams per kilogram. Compare this range with the 9.8 milligrams of iron per kilogram of the transgenic iron-fortified rice IR68144-2B-2-2-3, developed at IRRI at enormous expense.

Certain landraces may have medicinal uses. Ayurveda, the traditional Indian system of medicine, recommends Nyavara rice from Kerala to help treat a class of neurological disorders. Along with my co-workers, I am examining its chemistry and also hope to study its efficacy for such use. Another medicinal rice, Garib-sal from West Bengal, was prescribed in traditional medicine for treatment of gastroenteric infections. In a 2017 paper in ACS Sustainable Chemistry and Engineering, my collaborators and I documented the bioaccumulation of silver in Garib-sal grains to the extent of 15 parts per million. Silver nanoparticles kill pathogenic bacteria, according to a 2017 study in Chemistry Letters, so this rice might help fight human gut pathogens. A plethora of such medicinal rice varieties awaits laboratory and clinical testing.

Aesthetics is yet another value that indigenous farmers cherish, cultivating certain landraces simply for their beautiful colors or patterns: gold, brown, purple and black furrows on yellow hulls, purple apexes, black awns, and so on. Many in eastern India take pride in the beauty of the winglike extensions of the sterile lemma in Moynatundi and Ramigali rice. Aromatic varieties are associated with religious ceremonies and cultural festivals in all rice-growing cultures. When these types of rice disappear from fields, numerous culinary delicacies are no more, and the associated ceremonies lose their cultural and symbolic significance. Basudha’s collection of 195 aromatic rice landraces has helped revive many evanescent local food cultures and traditional ceremonies.

The complexity of ecological interactions has resulted in another set of rice varieties. Smallholding farmers of West Bengal and Jharkhand prefer varieties with long and strong awns (spine-like projections at the end of the hull), which deter grazing by cattle and goats. Indigenous farmers also prefer landraces with erect flag leaves because grain-eating birds cannot perch on them.

Interestingly, some farmers in Odisha grow a combination of awned and awnless varieties on their farms, regardless of any direct benefits. Other rare varieties with no obvious use possess purple stems and leaves. Indeed, South Asian tradition appears to deem biodiversity, at both the genetic and the species level, as so essential to agriculture that it was enshrined in certain religious rituals. For example, some wild relatives of cultivated rice, such as Buno dhan (Oryza rufipogon) and Uri dhan (Hygroryza asiatica), are associated with local Hindu rites and maintained on many farms in West Bengal and its neighboring state, Jharkhand. Such wild gene pools are becoming ever more important as a source of unusual traits that can be incorporated, as required, into existing cultivars. Further, the presence in rice fields of certain trees such as neem (Azadirachta indica), whose leaves serve as a natural pesticide, and of predators such as the owl has been considered auspicious.

Saving Farmers

Given the failure of modern agricultural research to provide marginal farmers with any reliable germ lines of rice, a large collection of folk rice varieties, with their fine-tuned adaptations to adverse conditions, is our best bet. Convinced by the superior yield stability of the landraces, more than 2,000 farmers in Odisha, Andhra Pradesh, West Bengal, Karnataka, Kerala and Maharashtra have adopted several folk rice varieties from Vrihi and abandoned cultivation of HYVs.

When Cyclone Aila hit the Sundarbans coast of West Bengal and Bangladesh in May 2009, it killed almost 350 people and destroyed the homes of more than a million. A storm surge inundated fields with seawater and left them salinated—which meant that quite apart from the immediate devastation, the food security of the region was likely to suffer long-term damage. We distributed a small amount of seeds from the Vrihi seed bank’s repertoire of traditional salinity-tolerant landraces, such as Lal Getu, Nona bokra and Talmugur, among a few farmers on island villages of the Sundarbans. These were the only rice varieties that yielded a sizable amount of grain on the salinated farms in that disastrous season. Similarly, in 1999 several folk varieties such as Jabra, Rani kajal and Lakshmi dighal ensured rice production for southern Bengal farmers after a flash flood of the Hugli River. In 2010 Bhutmuri, Kalo gorah, Kelas and Rangi rescued many indigenous farmers in the western district of Puruliya when delayed arrival of monsoon rains caused a severe drought.

Such disasters prove, time and again, that the long-term sustainability of rice farming depends crucially on the restoration of traditional farming practices based on biodiversity and use of the full diversity of crop varieties that have survived the onslaught of industrial farming.