The discovery began as a suspected disease. Émer Hickey was gardening with her mother in Kinsale, on the Cork coast of Ireland, when they pulled up a pea plant and found its roots studded with pale, wart-like lumps. It looked sick. Émer, then in her early teens, did the thing that separates a science story from a compost story: she took the plant to school and asked her teacher what was wrong with it.

Nothing, was the answer. The warts were nodules, and inside them lived rhizobium, a soil bacterium in the nitrogen-fixing family called diazotrophs, which legumes like peas deliberately house in exchange for fertilizer. The bacteria pull nitrogen from the air and hand it to the plant as ammonia; the plant pays in sugar. It is one of agriculture’s oldest partnerships, and one of its most frustrating exclusions, because the world’s staple cereals, wheat, barley, rice, cannot form it.

Émer told her friends Ciara Judge and Sophie Healy-Thow, whose geography class happened to be studying the world food crisis in the shadow of the 2011 Horn of Africa famine, and the three of them landed on a question with billion-dollar aid programs attached: if cereals can’t host these bacteria, could the bacteria still help cereals?

The spare-room laboratory

The advice from grown-up science was discouraging. Rhizobia’s benefits were understood to belong to legumes; many people told them the bacteria would have no impact on cereal crops. The girls, aged around 14 when they started, decided the objection had never actually been tested at their scale, and set out to test it at a scale nobody expected.

Working from home, with a bedroom in the Judge house converted into a laboratory, incubation racks, hand-labeled samples, a homemade regime of controls, they soaked cereal seeds in rhizobium cultures and measured everything: time to germination, germination rate, seedling growth, dry mass. Batch after batch, season after season, for three years, they ran the experiment across barley, oats and wheat, ultimately testing some 13,000 seeds, plus thousands more in an outdoor field trial, and logging on the order of 120,000 individual measurements by hand into their notebooks and spreadsheets.

The seeds ignored the expert consensus. Treated barley and oats germinated dramatically faster and more reliably, with germination rates improved by up to 50 percent, and the head start carried through: in their growth trials the treated cereals put on substantially more dry mass, with increases running as high as 74 percent. The bacteria could not give barley a legume’s nodules, but something in the association, delivered at the seed, was waking the grain up early and sending it into the world stronger.

The scoreboard

Irish science noticed first. The project won the BT Young Scientist of the Year in 2013, making the trio the first group of girls ever to take Ireland’s premier young-science title, then the EU Young Scientist contest. In September 2014 came the summit: at Google’s headquarters in California, “Combating the Global Food Crisis: Diazotroph Bacteria as a Cereal Crop Growth Promoter” was named grand prize winner of the Google Science Fair, beating thousands of entries worldwide. The prize haul included scholarship money and a trip to the Galápagos; Time magazine put Ciara Judge among the world’s most influential teens. They were 16 and 17, and they had started at 14, on a plant most gardeners would have thrown away.

The relevance of the result is not subtle. Germination is one of farming’s silent taxes; seeds that sprout slowly or not at all cost yield before the season begins, and a faster, more uniform start matters most exactly where the growing seasons are shortest and the margins thinnest. A seed treatment based on naturally occurring soil bacteria, needing no synthetic chemistry, is the kind of low-cost tool that travels well to smallholder agriculture, which is why the girls framed the project around food security from the first poster. They talked afterward about patents and commercialization, and the wider science has moved their direction since: engineering cereals and microbes to extend nitrogen-fixing partnerships beyond legumes is now one of agricultural research’s most funded ambitions.

The honest footnote is the usual one for prodigy science: a bedroom study, however heroically sized, is not a peer-reviewed field program, effects of “up to” 50 and 74 percent are ceilings rather than averages, and no commercial product yet traces to the Kinsale data. What stands is harder to discount. Three teenagers were told a biological door was closed, checked the literature, found that nobody had pushed on it with 13,000 seeds and three years of patience, and pushed. The pea plant’s warts turned out to be the least diseased thing in the story. The ailment was the assumption, and the cure, as it often is, was a child asking what, exactly, is wrong with this thing you were about to throw away.