LAST SPRING, the U.S. Forest Service sprayed a valley of Oregon timberland with a mixture of two herbicides. Within two weeks, five of the seven pregnant women in the area had suffered miscarriages. In the ensuing weeks, virtually every one of the 50-odd households in the valley reported new health problems, including respiratory and intestinal problems, uterine hemorrhaging, high fevers and skin rashes. Spraying in this same valley a few years ago with the related herbicide 2,4,5-T led to a rash of miscarriages, an intensive study by the EPA and the subsequent suspension of that chemical while hearings are held on a permanent ban. For the residents of the area, it feels like a recurrent nightmare.

For regulators in Washington, it seems a puzzle. One of the two herbicides, 2,4-D, is among the most widely used in the world and has been in use for many years. Forest Service employees, including women, who apply the stuff get huge exposures without incident. There is little prior evidence showing a correlation between 2,4-D and the kind of health problems found in Oregon. However, it is notoriously difficult to collect such data from farmers and other residents of dispersed rural communities, and there are laboratory test results showing a relationship between the two herbicides and several illnesses, including cancer. Finally, the economic consequences of suspending the use of 2,4-D could be very severe.In short, for every "on the one hand" there is an "on the other."

In the past decade, the doctrine of informed consent has become a central tenet of proper medical practice. Congress has passed numerous laws protecting and strengthening the citizen's right to know what the government is doing to him and what it knows about him. Non-smokers have gotten militant about the effects other people's smoking has on their health, and no-smoking areas have been set aside for them. But the same kind of attention to the effects of toxic chemicals on unsuspecting third parties has not yet been exercised either by government or private industry. In most cases so far, serious damage has been suffered by many people before enough "solid information" has been collected on which to base a suspension, closing, restraining order or whatever. For anyone who lives near a chemicals plant, on the same groundwater system as land that gets sprayed, or who is down river from a fertilizer factory, or who eats treated fruits and vegetables -- in short, for everyone -- it's clear now that that's not good enough.

A big and urgent effort needs to be made to improve our knowledge of epidemiology and toxicology and to increase the government's abilities in both specialties. The resources needed to thoroughly study the thousands of chemicals already in use as well as to stay abreast of the thousands of new ones that are introduced each year far outstrip what is currently available. Meanwhile, the experiences in Oregon (and in many other places) strongly suggest that if regulators must err, they should do so on the side of caution.