For 2,500 years, doctors unleashed a single, gory cure-all on nearly every ailment known to mankind a good bleeding. Confronted with a fever, rattly cough, intermittent seizures, heart disease, or even mental illness, doctors bled their patients, sometimes until they passed out.

The common practice of bloodletting weakened and probably killed some patients including George Washington, who was bled of 2 to 3 quarts of blood after getting sick and died shortly thereafter. Undaunted, barber-surgeons and doctors continued bloodletting to the cusp of the 20th century, when it was finally consigned to the trash bin of discredited medicine.

Now, however, a discovery published last month in the journal Science suggests that bloodletting actually might have helped some patients and offers a tantalizing reason why. Microbiologists at the University of Chicago found that when infectious Staphylococcus aureus bacteria were deprived of the iron in red blood cells, they did not spread and cause disease in the body.

"In the earliest 20th century, some of the most respected physicians advocated using bloodletting and had honed it down to use at the beginning of a sudden onset with a harsh fever, which is very much the hallmark of a bacterial infection," said Dr. Tracey Rouault, an expert on iron metabolism who looked at the Chicago study and saw a connection to the painful ancient practice."Right at that point, you may be doing some good."

No one today is advocating a return to the lancet. But if Rouault is right, bloodletting may be joining the list of cures from the past, including maggots and leeches, proving to be of interest to modern doctors.

"If something rather invasive has a hold on therapy for 2,500 years," Rouault said,"there must be some incident when somebody benefited from it."

The scientists behind the bacterial study didn't set out to find a justification for bloodletting. They were trying to solve a mystery about bacteria: The cells need iron to grow and thrive, but doctors have never known which sources of iron in the body were the first choice. So a research team at the University of Chicago put bacteria on a low-iron diet and then set them loose in a dish, with the two most common forms of iron found in the body.

Researchers found the bacteria overwhelmingly preferred heme iron, which is stored within red blood cells, especially at the beginning of the infection."They blow open the red blood cells, which leads to the release of heme," said Eric Skaar, a microbiologist at Chicago who coauthored the study.

He and his colleagues found that when the bacteria were altered so they could no longer capture the iron stored in the red blood cells, they were unable to cause disease in mice.

Rouault said that when she first saw the study results, she was excited. The head of the section on human iron metabolism at the National Institutes of Health, she had been musing on the rationale for bloodletting,"because it's just so curious." The new study suggested an answer: Maybe it caused iron starvation in bacteria. She wrote a commentary drawing the connection, which appeared in the same issue of Science.

Other science historians, however, caution that Rouault's idea does not mean bloodletting was an effective therapy. The only cases in which it would have worked were bacterial infections a fraction of the total number of diseases treated with the technique over the centuries, at a time when doctors had no way to tell a bacterial infection from a tension headache.