AS only Nixon could go to China, perhaps only a Democratic administration can join a fight against one of the worst health laws on the books, the Delaney Clause. This provision of the Federal Food, Drug, and Cosmetic Act was passed in 1958. It bans any additive in any detectable levels in processed food if the additive has been shown to cause cancer in humans or laboratory animals.

The law seems to make sense. Who wants a cancer-causing substance added to their food? But its broadly stated mandate does not reflect today's scientific knowledge and our growing awareness of health risks.

The Delaney Clause is subject to a lawsuit by the Natural Resources Defense Council. NRDC charged, accurately, that the EPA was failing to enforce the Delaney Clause. It was, because EPA realized that it was an unworkable law that does more to threaten our nation's health than help it. However, a California judge ordered EPA to follow the law.

The EPA's response to the decision was twofold: It began to list the chemicals that would be banned if it were forced to follow the letter of the law and it asked Congress to amend the law to reflect current scientific understanding.

The scientific problems are many. Dr. Bruce Ames of the University of California at Berkeley helped pioneer our modern techniques of testing substances for cancer effects. He is now among the methods' chief critics. He cites two concerns.

The first has to do with the testing method of feeding populations of animals the maximum tolerable dose (MTD) of a chemical, the dose just below that which causes death. After exposure the animals are examined for cancerous growths.

The problem is that the "dose-response" curve is unlikely to be linear, that is, continuous high exposures are likely to be more dangerous than long-term low exposures. A substance's ability to cause cancer at 50 or 5000 times the average daily intake level may tell us nothing about its effect upon cancer risk at smaller, more likely doses. As a government scientific panel announced in the Federal Register, "the implicit assumptions underlying extrapolations from the MTD . . . do not appear to be valid."

The second problem is transferring lab results on mice to human populations. Mice and rats are more closely related to one another than to humans. However, one will often test negative while the other will test positive for the same chemical.

Dr. Ames also points out that it is not only man-made substances that cause cancer. Using the same disputed techniques, he found a variety of other naturally occurring chemicals, present in substances from peanut butter to broccoli, that also cause cancer in test animals.

In contrast to what was believed when Delaney was passed, experts today believe that, aside from workers who are exposed to massive doses, not more than 1-3 percent of cancer cases can be attributed to man-made chemicals. The vast majority are due to genetics and lifestyle choices such as smoking, alcohol consumption, or fatty foods.

Of course, to those who campaign for public health, 3 percent is still too much. But they lose sight of the fact that relative, not absolute, risk is the important measure. Most chemicals are present for a reason. They help to control pests and bacteria. They lengthen shelf life. All act in ways that reduce the price of food and often the real danger associated with eating it. The result of strict enforcement of Delaney would likely be greater scarcity of some foods, price increases, and increased likelihood of contamination.

The environment might also suffer as more harmful substances are used. In Wisconsin, for example, farmers would likely switch to copper- or tin-based compounds to protect their potato crops from blight. Ultimately, this practice would poison their soils with the accumulation of heavy metals.

So by banning pesticides, we may increase other risks, shifting dangers elsewhere. But we cannot eliminate all risks, only choose intelligently among them. This requires a careful analysis of relative risks, not legislation where emotions trump science.

EPA's suggestion to relax Delaney is likely to face a tough battle in Congress. Few senators and representatives want to be portrayed as voting for cancer. Cancer is complex, emotional, and scary. And certainly we should seek to lower cancer risks when reasonable. But it is relative risks, what happens on the margins, that matter. To promote a safe and affordable food supply, lets be alert to environmental hyperbole and posturing while respecting medical science and the economics of risk reduction.