"The participants who were at the lowest risk for stroke were men who consumed one or two drinks on three or four days a week," said Dr. Kenneth Mukamal, the lead author of the study, an assistant professor of medicine at Harvard Medical School and internist at Beth Israel Deaconess Medical Center in Boston. "The importance of drinking pattern for stroke risk parallels our previous findings among this same group of men regarding alcohol consumption and the risk of developing diabetes and heart disease," Mukamal said. "Among all three types of disease, the lowest risk seems to occur when consumption is limited to one, or at most two, drinks, approximately every other day, with little benefit shown above three to four drinking days a week."

Data for the study came from the Harvard School of Public Health's Health Professionals Follow-up Study, which began in 1986 and did evaluations of male participants ranging in age from 40 to 70 every four years through 2000. Each participant completed a detailed questionnaire on diet and medical history, including alcohol consumption, at each interval.

During the course of the study, researchers confirmed 412 cases of ischemic -- clot-induced -- stroke among the group. Nearly 700,000 Americans suffer this type of stroke each year; such "brain attacks" account for about 80% of all strokes, with burst vessels cutting off blood supply to the brain.

The study showed that men whose average alcohol intake was three or more drinks a day had a 42% higher risk of ischemic stroke -- and particularly blood clots that moved to the arteries of the brain from elsewhere in the body -- than did abstainers.

While the study showed that average intake of lower amounts of alcohol did not significantly raise or lower the risk of stroke, when frequency of drinking was taken into account, it showed that light and moderate drinkers imbibing three or four days a week had a 32% lower stroke risk than nondrinkers.

Even so, Mukamal does not feel the study shows drinking has as strong an effect in guarding against stroke as against heart disease.

"While there does appear to be a small window for which light drinking is associated with lower risk," Mukamal said, "it's important to note that this window is smaller than it is for heart disease, and therefore, you cannot simply extrapolate between the two."

The U.S. government defines a "standard drink" as a 12-ounce beer, a 5-ounce glass of wine or a 1.5-ounce shot of distilled spirits -- each of which contains about one-half an ounce of alcohol.

Generally, it did not much matter what the men were drinking, although the small share who drank one or more glasses of red wine daily were at 46% lower risk.

"This is curious, because among this population of men, red wine is not linked to a lower risk of heart disease any more than any other type of alcohol, so it's unclear why this would be the case with stroke," Mukamal said, adding that more research in this area needs to be done.

Overall, the researchers feeling their findings "support current public health recommendations that men should consume less than two drinks per day."