PARSNIPS CAUSE cancer. Coffee beans may help prevent cancer. Meats will raise your blood pressure.

Findings like these, which used to sound like the ravings of the food faddist fringe, now regularly show up in respected medical journals. They're bits of evidence in a new puzzle intriguing many scientists: What are the physiological changes caused by many common foods, and how do those changes contribute to chronic diseases?

It's a fascinating and serious search, logical when you think about it. After all, many roots and herbs, used medicinally for years and sometimes converted into modern pharmaceuticals, have potent biological properties. Why should the foods we eat be any different? They do, undoubtedly, contain age-old secrets to our health, mysterious poisons -- no less lethal because they take so long to work -- or even antidotes to some of our modern environmental poisons. The findings are fragmented, sometimes contradictory and confusing. But they are pushing us toward an ever more sophisticated view of food toxicity and perhaps some profound dietary changes.

Here are only a few of the findings that have circulated in scientific journals and at meetings in the last few weeks:

* The case against parsnips. (Science magazine, Aug. 21) U.S. Department of Agriculture scientists in Texas determined that parsnip roots -- the part we eat -- contain high amounts of potent toxic chemicals called psoralens. These psoralens are known to cause cancer and mutations (also indicative of cancer) in laboratory animals. Cooking did not destroy the chemicals, nor did the peeling, although peeling reduced the presence of the chemicals by 30 percent.

It's clear these investigators don't think parsnips are good for you. They conclude that such foods "may present some toxicological risk to man," but like other reputable scientists, they call for human epidemiological studies to be sure -- in other words, a survey of the fates of those who eat parsnips. On this, as with saccharin, one can enter into that scientific thicket of how many parsnips one would have to eat to develop cancer. But as with saccharin, that is beside the point. The question is whether one should knowingly consume identified carcinogens, whether they be natural or artificial, especially if they are fairly potent.

Such chemicals do not on their own cause cancer, but they do contribute to the carcinogenic burden we all share from the air, water, food and soil. The implicit scientific question is: Why add one more carcinogen to put additional strain on the body's immune system against cancer? My vote -- if anyone asked -- would be against parsnips.

Incidentally, the scientists also tested carrots -- a close kin of parsnips, and detected little or none of the toxic chemicals. Not surprising, perhaps; other reputable studies have shown that carrots may help protect against cancer, especially lung cancer, possibly because of their high vitamin A or carotene content.

* The case for coffee. (A paper at the American Chemical Society conference in New York City, Aug. 27) Coffee has taken a lot of scientific trouncing lately. The most recent was a study in the prestigious New England Journal of Medicine linking coffee drinking to pancreas cancer in humans. So, this new report is somewhat of a surprise. The very respected researchers, Dr. Luke K. T. Lam and Dr. Lee Wattenberg of the University of Minnesota, found that coffee beans, notably green coffee beans, retarded the development of cancer in laboratory animals. The scientists exposed mice to a powerful cancer-causing chemical, and then fed some of the mice a diet of 20 percent powdered green coffee beans. Mice getting the coffee beans developed only half as many breast cancers as those not getting the coffee.

The scientists believe ingredients in the coffee beans acted as an antidote to the cancer-causing chemical by stimulating the enzyme system in the animals to combat the cancer. The investigators, in fact, identified two such chemicals in the coffee. They also found that roasted coffee beans and instant coffee, both regular and decaffeinated, triggered the cancer-fighting defenses of the mice -- but at only about half the rate as the green coffee beans.

Coffee lovers may rejoice, but the researchers are extremely cautious. Lam, in an interview, stressed that there's no assurance human enzyme immune systems work the same way as those in the experimental animals. He said the study did not refute the coffee-pancreas cancer findings (although that report has been highly criticized by other authorities).

Lam definitely does not recommend drinking lots of coffee to try to prevent cancer -- sensible advice, since this one cheery bit of evidence in favor of coffee is far overshadowed by the documented detrimental effects of excessive coffee drinking, including neurological changes, promotion of cystic breasts in women and, most important, dangers to the fetus. Still, the pro-coffee bean study is important, the researchers believe, because it may lead to the extraction of anti-cancer chemicals that one day might even be used as drugs to block the development of cancer.

* Meat versus vegetables. (Journal of the American Medical Association, Aug. 7) A team of well-known heart researchers, including Dr. William P. Castelli, author of the famed Framingham heart study, induced 21 strict vegetarians to eat about nine ounces of meat a day in place of the same number of calories they usually ate in grains and vegetables. Two effects were noticed during the eight-week period. The group's systolic (top number) blood pressure went up by about 3 percent and came back down again when meat was abolished from the diet. Also, the total amount of cholesterol in the blood rose by an average 19 percent, from a low of 140 milligrams to a high of 166 milligrams.

However, the amount of so-called "good cholesterol" in the blood, called high-density lipoproteins (HDL's) was not affected by eating meat. But the ratio of HDL's to total cholesterol did change, the scientists say, to the extent that the vegetarians were lifted out of the low-risk category for heart attacks into the high-risk category.

This study tends to confirm what some have been saying for years: that meat causes biological changes, especially in the fatty content of the blood, predisposing one to heart disease. The study (although it has certain flaws, according to an editorial comment in JAMA) also confirms any number of other studies which show that vegetarians generally have lower rates of chronic diseases, including heart disease and cancer.

Unquestionably, the causes of chronic diseases are multiple, complex and interrelated. And no one knows how the numerous chemical constitutents in foods interact with each other and the body to promote or retard disease. But the medical literature provides clues, which we can act on along the way, so that the final solution to the puzzle, once it comes, will be no surprise.