Some of the most compelling evidence linking a widely used pesticide to developmental problems in children stems from what scientists call a “natural” experiment.

Though in this case, there was nothing natural about it.

Chlorpyrifos (pronounced klor-PYE-ruh-fahs) had been used to control bugs in homes and fields for decades when researchers at Columbia University began studying the effects of pollutants on pregnant mothers from low-income neighborhoods. Two years into their study, the pesticide was removed from store shelves and banned from home use, because animal research had found it caused brain damage in baby rats.

Pesticide levels dropped in the cord blood of many newborns joining the study. Scientists soon discovered that those with comparatively higher levels weighed less at birth and at ages 2 and 3, and were more likely to experience persistent developmental delays, including hyperactivity and cognitive, motor and attention problems. By age 7, they had lower IQ scores.

The Columbia study did not prove definitively that the pesticide had caused the children’s developmental problems, but it did find a dose-response effect: The higher a child’s exposure to the chemical, the stronger the negative effects.

That study was one of many. Decades of research into the effects of chlorpyrifos strongly suggests that exposure at even low levels may threaten children. A few years ago, scientists at the Environmental Protection Agency concluded that it should be banned altogether.

Yet chlorpyrifos is still widely used in agriculture and routinely sprayed on crops like apples, oranges, strawberries and broccoli. Whether it remains available may become an early test of the Trump administration’s determination to pare back environmental regulations frowned on by the industry and to retreat from food-safety laws, possibly provoking another clash with the courts.