In a hot, dry field near a place called Humpty Doo in Australia’s Northern Territory, scientists are racing to begin an experiment that could determine the future of the world’s most popular fruit, the lowly banana.

Dodging the occasional crocodile, researchers will soon place into the soil thousands of small plants that they hope will produce standard Cavendish bananas — the nicely curved, yellow variety representing 99 percent of all bananas sold in the United States. But in this case, the plants have been modified with genes from a different banana variety.

An insidious fungus known as fusarium wilt has wiped out tens of thousands of acres of Cavendish plantations in Australia and Southeast Asia over the past decade. And the fungus recently gained a foothold in Africa and the Middle East, hitching a ride on the boots of workers helping to establish new plantations. Scientists say Latin America, the source of virtually all the bananas eaten in the United States, is next.

No other variety of banana combines the sweetness and suitability for packing and export of the Cavendish. If the Humpty Doo experiment — or simultaneous efforts with conventional breeding techniques — don’t bring positive results, scientists say we could be looking at a future where bananas all but disappear from store shelves.

“These recent outbreaks confirmed that this thing does move,” said plant pathologist Randy Ploetz of the University of Florida, who first identified the fungus in 1989 in samples from Taiwan. Ever since, banana farmers have been trying to escape the effects of fusarium wilt, also known as Panama disease Tropical Race 4, or TR4. Fungicides and fumigants are useless against it. It’s extremely contagious, and it can lie dormant for decades, tricking farmers into thinking they’ve eliminated the pathogen, only to find plants rotting from the inside.

Soon after TR4 was identified, banana farmers had reported that a subspecies of the Musa acuminata variety of sweet bananas, which grows in the wild across Malaysia and Indonesia, was “growing happily in plantations devastated by TR4,” said James Dale, a professor of biotechnology at Queensland University of Technology in Australia.

It took years to isolate the gene responsible for the resistance. But Dale’s lab had a breakthrough: It identified candidate genes worth testing. Over three more years of painstaking work, Dale inserted genes from the M. acuminata subspecies into cells from a Cavendish, developing them first in tiny test tubes, then growing whole plants.

Dale’s project may be the best hope science now has for making the Cavendish resistant to TR4 without eliminating taste, texture and other characteristics that make it so appealing and commercially successful.

Ploetz is optimistic about Dale’s trial, but he’s not peeling his bananas before they’re ripe. He thinks Dale needs “to test this new [genetically modified fruit] in different environments and see the effects on yield” and other factors, he said.