On this page
Rice feeds more people than any other crop, and for much of the twentieth century Asia struggled to grow enough of it. Gurdev Khush, who grew up on a farm in northern India, spent over thirty years breeding rice that grew faster, fought off pests and yielded more. This post covers his path, what made IR36 different, and a worked example on the numbers plant breeders care about.
Who is Gurdev Khush?
The World Food Prize says Khush grew up on a wheat farm in northern India and graduated from Punjab Agricultural University in 1955. He went to England and worked in a factory to save money for travel to the USA, and in 1957 was accepted at the University of California, Davis, where he studied with the geneticists Ledyard Stebbins and Charles Rick. After his PhD in 1960 he spent seven years studying the tomato genome with Rick.
In 1967 the International Rice Research Institute (IRRI) in the Philippines recruited him to breed high-yielding rice with short growth duration, resistance to diseases and insects, and good grain quality. Within five years he headed IRRI's plant breeding department. He retired from IRRI in 2002 and returned to UC Davis as an adjunct professor.
What was IR8, and what was wrong with it?
The World Food Prize page tells the story. IRRI's IR8, selected by Henry Beachell in the 1960s, was short and sturdy, responded to fertiliser and more than doubled the 1 to 2 tons per hectare of traditional varieties. But it had low resistance to some pests and diseases, and many rice eaters did not like how it cooked.
What made IR36 different?
In 1976, the same page says, Khush released IR36, bred by crossing IR8 with 13 parent varieties from six countries. It was:
resistant to several major insect pests and diseases
quick to mature, in 105 days, compared with 130 days for IR8 and 150 to 170 days for traditional types
slender-grained, the kind many countries prefer
About 11 million hectares were planted with IR36 in the 1980s, and the World Food Prize calls it the most widely planted food crop variety ever grown. Khush followed it with IR64 and IR72. The Prize says the 300-plus improved varieties from his programme went to rice-growing countries across Asia, Africa and Latin America.
The Wolf Prize in Agriculture, which he received in 2000, highlights the research behind this: genes for resistance to the brown planthopper, green leafhopper, whitebacked planthopper and gall midge, a complete set of trisomic lines for mapping rice chromosomes, and, with colleagues, the first molecular genetic map of rice.
Worked example: reading rice numbers like a breeder
All numbers in steps 1 and 2 come from the World Food Prize page.
Time saved. IR36 matured 130 - 105 = 25 days sooner than IR8, and 45 to 65 days sooner than traditional rice. A shorter crop frees the field earlier, which can make room for another crop in the year.
Production growth. World rice production rose from 257 million tons in 1966 to 686 million tons in 2010. The ratio is 686 / 257 = 2.67. Over 44 years, the average yearly growth r satisfies (1 + r)^44 = 2.67, so 1 + r = 2.67^(1/44) = 1.0226, about 2.3 per cent a year.
Harvest index. Breeders split grain yield into total plant biomass times harvest index, the fraction of biomass that ends up as grain. A 2000 study by Peng, Khush and colleagues in Crop Science grew IRRI varieties released since 1966 side by side and found yields rising about 1 per cent a year with year of release. Gains before 1980 came mainly from a higher harvest index; later gains came from more biomass. With sample numbers: 18 tons of biomass per hectare at a harvest index of 0.5 gives 18 x 0.5 = 9 tons of grain.
The same study warns that the newest varieties gave 9 to 10 tons per hectare, about what IR8 gave at the same sites in its early years, so the yearly trend may not mean a higher ceiling on yield. That is a useful habit: ask what a trend actually measures.
Common mistakes:
Averaging growth by division. 2.67 times in 44 years is not 167 / 44 = 3.8 per cent a year. Growth compounds, so take the 44th root.
Confusing yield with production. Yield is per hectare; production is yield times area.
If compound growth and roots of numbers are the step that trips you, the Fermi Problem Library has worked problems with a step-by-step solver that explains the reasoning rather than printing an answer.
How does his work connect to India's Green Revolution in wheat?
The same ideas, short sturdy plants that respond to fertiliser plus bred-in disease resistance, drove the wheat story. Read our posts on B. P. Pal's rust-resistant wheat, Norman Borlaug and M. S. Swaminathan.
What can a student take from Khush's way of working?
Fix the flaws of a success. IR8 was a breakthrough, and Khush still listed what was wrong with it and bred those faults out.
Combine basic and applied work. Chromosome mapping and pest genetics fed directly into varieties farmers planted.
Check what the numbers mean. His own study questioned whether rising yields reflected real genetic gain.
Try it now: if a variety has 16 tons of biomass per hectare and a harvest index of 0.45, what is its grain yield?
