But even Sir Christopher Andrewes, the British virologist who co-discovered the influenza virus in 1933, once observed, "I have always had my doubts about this."

Andrewes, who died in 1988, argued that the winter crowding explanation defies common sense, since office and factory workers are stuck indoors year-round. And summertime flu epidemics remain rare despite the prevalence of cruise ships, airplanes, and subways packed with international travelers, other skeptics of the theory note.

"There are plenty of things that don't fit," says the CDC's Dowell.

The latest theory to generate a buzz centers on sunlight.

In 1981, a British general practitioner named R. Edgar Hope-Simpson published the first paper documenting a link between influenza epidemics in the northern hemisphere and the winter solstice.

The solstice - which arrives on Dec. 21 this year - is commonly identified with the start of winter and is the shortest day of the year.

Hope-Simpson, who had no formal training in epidemiology, realized that influenza infections tended to jump just before and after the solstice. Solar radiation, he surmised, triggered some sort of "seasonal stimulus" that affected the flu virus, its human host or both.

Although he had no idea what that seasonal stimulus was, Hope-Simpson believed that solving the mystery "would provide the key to understanding most of the influenza problems confronting us."

"His work was basically ignored," says Dr. John Cannell, a psychiatrist at the Atascadero State Hospital in California.

In their forthcoming paper in Epidemiology and Infection, Cannell and colleagues from Harvard University, the National Institutes of Health and Boston University propose that Hope-Simpson's seasonal stimulus could be vitamin D.

The tip-off, says Cannell, came in April 2005, when an outbreak of influenza swept through Atascadero, a maximum-security facility for the criminally insane 200 miles south of San Francisco.

"The ward below me got it, then the ward to my right, to my left and the one across the hall," he recalls.

Cannell's 32 patients, however, were spared.

As far as he could tell, there was only one glaring difference between his patients and those who came down with flu: Cannell's patients were taking high daily doses of vitamin D.

Manufactured by the skin in response to solar ultraviolet B radiation (UVB), vitamin D is the only vitamin made naturally in the body.

On sunny days, a fair-skinned person can crank out the equivalent of 20,000 international units (IU) of the vitamin in 15 minutes, says Cannell. One cup of fortified milk, by contrast, contains fewer than 100 IU.

But vitamin D production plummets during winter because people spend less time outdoors and because UVB radiation has a harder time penetrating the atmosphere in that season, especially at middle to high latitudes.

As a result, health experts in recent years have warned that many Americans may not be getting as much vitamin D as they need. One ominous sign: a recent resurgence in cases of rickets, a bone disorder caused by a lack of the vitamin.

In their new paper, which draws together strands from more than seven decades of vitamin and flu research, Cannell and his colleagues argue that vitamin D stimulates production of a natural infection-fighting substance in the body called cathelicidin.