The virus that causes AIDS has been discovered in the teardrops of a patient suffering from the disease, according to researchers at the National Institutes of Health.

The discovery is the first evidence that the virus is found in eye fluid, and it raises new questions about whether infection could be transmitted through direct contact with the tears of AIDS patients, particularly by eye doctors and their patients, as well as through contact with the medical instruments used in eye examinations.

There is no evidence that such transmission has ever taken place, said scientists from the National Eye Institute, the National Cancer Institute and NIH's Clinical Center.

But sources familiar with the new findings suggest that transmission of the HTLV-III virus -- the virus found to cause acquired immune deficiency syndrome as well as other illnesses -- could, in theory, occur through repeated contact with the tears of AIDS patients or others carrying the virus.

Finding the virus in tears is significant in understanding the disease, said Dr. Robert Gallo, head of NCI's laboratory of tumor cell biology. "It indicates that it is present increasingly in more body fluids than we or anybody else originally thought. It has been found not only in blood cells and lymph nodes, but also free in the blood plasma, in semen, saliva and now tears," he said.

Because frequent contact with tears occurs during eye examinations, government physicians are preparing recommendations urging that precautions be taken to minimize direct contact with the tears of AIDS patients, including contact during routine medical procedures such as testing for glaucoma and fitting soft contact lenses.

Such precautions might include the use of gloves by medical personnel during eye examinations and using bleach to sterilize equipment used to examine these patients' eyes.

The findings from the study have not been publicized, but have quietly begun to circulate in medical and government circles. A scientific paper has been submitted to the British medical journal Lancet by several authors, including Dr. Leslie S. Fujikawa of the eye institute's clinical branch and Dr. S. Zaki Salahuddin of NCI's laboratory of tumor cell biology.

Dr. Carl Kupfer, director of the National Eye Institute, declined to comment on the findings yesterday, saying, "I think it's vitally important that the professionals have this information before the lay public has it."

A key question is whether precautions should be taken in treating only known AIDS patients, those with symptoms of a less severe form of the disease known as AIDS-related complex, those at risk of getting the disease or all patients.

"We haven't made an official recommendation yet," said researcher Fujikawa. "My feeling is that until we know more, it's better to be safe than sorry. I believe precautions should be taken . . . . At the moment, we're mainly concentrating on patients that we know or suspect may be at risk for AIDS."

AIDS, which can destroy the body's immune system, is considered to be spread largely through sexual contact. Other transmission occurs through contact with blood through contaminated needles or transfusions. It has also been shown to be transferred from infected mothers to their unborn children.

AIDS experts have emphasized repeatedly that the virus does not appear to be spread through casual contact, but rather through intimate exchange of body fluids.

NCI scientist Salahuddin, who isolated the virus from tears, said that there is no evidence that AIDS infection has actually been transmitted among humans through contact with tears and that the chances of this occurring appeared extremely small. He said the virus is fragile and does not survive well outside the body fluids.

Salahuddin's laboratory detected evidence of AIDS virus in a few drops of tears taken from the eye of a 33-year-old woman with AIDS. She had had two bouts with a rare pneumonia associated with AIDS, but was described as "ambulatory."

The scientists also found suspicious, lower-level virus activity in three other AIDS patients tested. Two others were negative, as was a patient with a less severe AIDS-related illness. Five tear samples from healthy NIH volunteers were negative.