The poison known as aflatoxin that contaminates food supplies can delay children’s development and stunt their growth. It also causes liver cancer, and exposure to high enough concentrations can be fatal. The toxin is produced by a common mold, Aspergillus flavus, that grows on crops from rice and cereals to corn and nuts.

But only some of these molds produce the toxin. Now, researchers have shown that insects spur A. flavus to make aflatoxin, suggesting ways to keep it out of the world’s food supply.

“All in all this makes sense to me, and it’s what I would predict,” says Nancy Keller, a microbiologist at the University of Wisconsin in Madison who was not involved in the study. “It’s nice to have a solid piece of [research] out on this.”

In addition to its health effects, aflatoxin also makes its way to the farm animals that eat these crops, causing an estimated $270 million in agricultural losses per year in the United States alone. In developing countries, the cost is higher.

The toxin likely costs the fungus, too, in energy and nutrients. But because more than two-thirds of A. flavus molds produce aflatoxin, researchers reason the poison must help out the fungus in some way.

To find out why only some A. flavus make the aflatoxin, Mickey Drott, a plant pathologist at Cornell University, and colleagues looked to fruit flies. The flies and fungi use the same plants as breeding grounds and eat the same foods. The fly larvae also occasionally dine on the fungus. So, the researchers thought the insect might prompt the mold to make aflatoxin to protect itself and its food from the bugs.

In their initial experiments, Drott and colleagues confirmed that the poison appears to protect the mold against the insects: When they added aflatoxin to the fruit fly larvae’s food, the maggots died and the fungus flourished. But that fungal growth only happened when the larvae were around; when the larvae were absent, the fungus didn’t grow. Toxic versions of the fungus also grew more when the bugs were around than when they weren’t, the researchers report today in the Proceedings of the Royal Society B. And the fungus made more toxins when larvae were around than when they weren’t. All of this suggests that, when bugs are present, aflatoxin is, too, the team concludes.

Yet, fruit flies rarely interact with the fungus in the wild—pests like the corn earworm caterpillar are much bigger threats—so it’s unclear how applicable these results are to the real world, says Marko Rohlfs, an evolutionary ecologist at the University of Bremen in Germany who was not involved in the study. “We urgently need model systems mimicking field conditions.”

Still, says Drott, his work suggests “this [interaction with insects] is a direction that people should start looking” in regard to toxin control strategies. Among other methods, current biocontrol strategies against the mold involve dousing fields with the nontoxic version of the fungus, which would hypothetically leave no place for toxic A. flavus to colonize. But, Drott and his team’s work demonstrates potential control methods should pay attention to crop insects as well.