Sometimes it takes science a while to get its act together. Take the curious case of the coffee bean.

This week a group of scientists from the University of Minnesota said a diet rich in coffee beans, which may contain a powerful anti-cancer factor, inhibits cancer growth in animal experiments.

Another group, from Harvard, defended its view that coffee drinking in humans may cause many cases of cancer of the pancreas. A Yale group says the Harvards are wrong.

Is coffee good for you? Or bad?

The truth is, no one knows. Science in progress is often confusing, with one study contradicting another. Things that work one way in animals often don't work the same way in humans.

But sometimes they do.

So have a cup of coffee, if that's your cup of tea, and we'll sort it all out.

First, coffee beans.

Dr. Lee Wattenberg of the University of Minnesota has been looking for substances in the diet that may block cancer. As his colleague, Dr. Luke K.T. Lam, puts it, "Human beings are being bombarded with carcinogens day in and day out," but most people don't get cancer.

So why not?

Some time ago Wattenberg found that something in cabbage and brussels sprouts inhibits cancer development in animal experiments.

Lam recently fed green coffee beans to a strain of rats that usually develops many breast cancers. Then he force-fed the rats a powerful cancer-causing chemical.

On a 10 percent coffee bean diet, 45 to 48 percent of the rats developed tumors, compared with 65 to 69 percent on an ordinary diet. On a 20 percent coffee bean diet, the number with tumors dropped to 25 percent.

Lam thinks a coffee bean ingredient called kahweol palmitate enhances an enzyme called glutathione S-transferase, which is one of nature's detoxifying agents in the body.

Tests on rat livers and small intestines, where the enzyme is active, showed, however, that roasted coffee beans had only half the enhancing effect of the green beans, and instant coffee a little less still.

Now, Harvard vs. Yale.

Harvard scientists under Dr. Brian MacMahon compared patients who had pancreas cancer with patients having other disorders. They found fewer non-coffee drinkers among the cancer patients than in the control groups.

This, they said, seemed to show that drinking up to two cups of coffee daily had increased the risk of pancreas cancer nearly twofold, and three or more cups, nearly threefold.

"I'm fairly convinced this is a real thing," said MacMahon, a highly respected epidemiologist, but "I would not give any advice to others yet."

The question is important. Cancer of the pancreas, a large abdominal organ that secretes digestive enzymes and insulin, is one of the most quickly fatal.

But a Yale group under Dr. Alvan Feinstein doesn't believe the Harvard work.

Feinstein and company say in the current Journal of the American Medical Association that the Harvard study was poorly designed and "our analysis" does not find coffee guilty in pancreas cancer. "People," said one of Feinstein's associates, "ought to have a much higher level of skepticism in results obtained in this manner."

MacMahon's rejoinder: "We found an association that is unlikely to be due to chance . . . . We're not saying the association exists. We're just saying it existed in our data" and "there should be further research."

Even if scientists decide someday that coffee really helps protect us from cancer, they will probably never say it is harmless.

Caffeine has been on trial for a possible role, also unsure, in causing some birth defects. Last year the Food and Drug Administration advised women who want to be "prudent" to stop or "minimize" coffee, tea and cola drinking during pregnancy.

Coffee has powerful gastrointestinal effects, which bother some people. Adults who consume large amounts, say 10 cups a day, may develop "caffeinism," or nervousness, irritability, agitation and headaches. Many doctors think children who drink too many colas may show the same symptoms.

Science is never simple.