VER the past century medicine has wrought so many miracles that progress has come to seem inexorable. A discovery is made, a cure is found. Next, please.

But the truth is that medical science, like all science, does not proceed from ignorance to enlightenment in a straight line. Indeed, it is more accurately described, in the words of the philosopher Miquel de Unamuno, as "a cemetery of dead ideas," with one seeming truth being thrown out for another that fits better with the latest research.

A particularly painful example may be occurring right now. For years, women over 40 years old were told that a yearly mammogram could find breast cancer early enough to save them from death.

This was medical dogma; it was the truth. But in the early 1990's, doubts grew about whether the test helped women in their 40's, and now some experts say they question whether it saves anyone.

"People don't like to hear this stuff," said Dr. Leslie Laufman, an oncologist at Hematology Oncology Consultants of Columbus, Ohio. "People cling to mammograms. They cling to the idea that there is something they can do to protect themselves. If we take away that as a security blanket, people turn to you and say, `So what am I supposed to do?' "

The mammogram story shows both the uncertainty of scientific progress and the slipperiness of scientific truth.

"There is a great ignorance of the scientific process," said Dr. Leon Gordis, a public health expert at Johns Hopkins University. All too often, science progresses in fits and starts, reexamining data, reinterpreting evidence — a path that can be hard to accept in medicine, when answers are needed now.

It happened recently with hormone replacement therapy for women after menopause. First it seemed that all women could benefit. Then scientists cautioned that estrogen might spur breast cancer. But, they said, it protects the heart, and more women die of heart disease than breast cancer. Now, it looks as if estrogen pills do not protect the heart. Estrogen does slow bone loss, and so it may help prevent fractures from osteoporosis. And it alleviates hot flashes. Who should take it? Experts disagree.

"Uncertainty is very painful," Dr. Gordis said. "The idea that science is not giving you certitude is very difficult for many people to accept."

The mammography argument, scientists say, is only taking place because the test's benefit was never that great in statistical terms. While the test can find tumors too small to feel, does finding and treating them that early save lives?

Some of the strongest evidence is from a study begun in the 1960's. It found that after 18 years, 153 out of 30,131 women who had mammograms had died of breast cancer, and 196 out of 30,565 women who did not have the test died of breast cancer. That is a 30 percent difference in breast cancer death rates — but it hinges on the medical histories of just 43 women. Questions about the design and conduct of this study have led some to doubt its conclusion. And similar questions have been raised about other mammography studies.

The question now is whether there are enough doubts about mammography's benefits and enough evidence of its risks that women should reconsider having them.

Some risks lie not in the test itself, but in the treatments — chemotherapy, surgery and radiation — when malignancies are found. But what if mammograms are picking up tumors that would never threaten a woman's life, and would otherwise never have been noticed — or treated? What if mammograms are finding tumors that could be treated just as well a few years later, forcing the woman to spend those years with the dreadful knowledge that she has breast cancer?

BUT many doctors say it is inappropriate for scientists to quibble about the fine points of evidence in the case of a devastating disease like breast cancer.

"Unfortunately, the people making these arguments are statisticians," said Dr. Maurie Markman, a specialist in gynecological cancer at the Cleveland Clinic.

"I'm not trying to say that statistics are not important," Dr. Markman said. "We can argue about how many angels can dance on the head of a pin. But these are real live patients, and it doesn't help anyone to go through this."

It defies belief, he said, that women should forgo the only test they now have that can find breast cancer early. "We surely find things that are small and those cancers are cured," he said.

BUT it is not just statistical arguments that cloud the picture, said Dr. Barron Lerner, an internist and historian of medicine at Columbia University's College of Physicians and Surgeons. There is also the image of cancer as a relentless enemy that will advance unless it is conquered. It is, he said, an image of war.

"What do you do in a war?" he said. "You search out the enemy, then you do what you can to surround and destroy the enemy. Cancer is the enemy and mammography is the tool we use to search it out."

To avoid asking women to decide whether they want to give up a weapon in this war, Dr. Lerner said, many doctors are acting as though the scientific questions are not serious.

Last week, for example, Dr. Stephen Brenner, an internist in private practice in New Haven, asked his female staff and a male resident a pointed question: was there anything to this new debate over whether mammography lowers the death rate from breast cancer?

The unanimous response, Dr. Brenner said, was that the entire argument was without foundation. "It's just a way for insurance companies to save money," they told him. In the meantime, he said, not a single patient has told him she questions the value of mammograms.

But ultimately, said Dr. Lerner, the data do cast doubt on long-held beliefs — and they do not replace them with definitive answers. "This is a story of the limits of evidence in science," he said.