The disease is autism, a heartbreaking illness that leaves children profoundly isolated from their parents and peers. The prevention is the combined measles-mumps-rubella vaccine, or MMR, which kids begin getting on their first birthday.

The theory is that MMR causes autism.

Around the world this year about 30 million people will die of measles, mostly children in countries that lack resources to vaccinate everyone. But it wasn't poverty that caused the measles to rip through Dublin this past winter. The outbreak, which sickened 1,000 kids and killed two of them, festered in a neighborhood where parents withheld their kids from "the jab" out of fears that the vaccine caused autism.

Measles is a highly contagious disease -- 120 kids died in the last U.S. outbreak a decade ago. And so, under pressure from Rep. Dan Burton, R-Ind., and many parents who believe that vaccines cause autism, the government recently agreed to spend a few million dollars investigating the vaccine theory.

Not all scientists are happy about that. "Having your national research agenda run by crazy people is not the best thing," says Dr. Patricia Rodier, an autism researcher at the University of Rochester. "The tragedy is that we're spending money that could help us find the real causes of autism." Still, with a disease like autism, which probably has many causes, all of them somewhat mysterious, anything is possible. If only to abate mistrust that might provoke families not to vaccinate their kids, the government had an obligation to fund the research.

"It would not be good to have a scare in which the public stops vaccinating," says Matthew Belmonte, a neuroscientist who teaches writing at MIT and studies autism at Harvard Medical School. "But even if there were absolutely no evidence for the vaccine theory, this is a democracy and the public has some right to dictate research."

That said, evidence for the vaccine theory is weak.

The most plausible explanation for the explosive increase in reports of autism is the 1987 publication of a new psychiatric diagnosis manual, which broadened the definition of the disease. As those guidelines, and general awareness of autism spread among doctors, speech therapists, pediatricians and parents, autism registries swelled.

Before this trend began, autism rates were much lower in the United States than in Europe. Missouri, for example, saw autism rates jump from 0.15 to 4.8 per 10,000 from 1988 to 1995 -- a 30-fold increase that's obviously an artifact of changed reporting criteria. In Japan, where the combined MMR vaccine is not given, a 1996 study showed a rate of 21 autistic children per 10,000.

The most compelling arguments for vaccines as the cause of the increase are the parents whose kids seemed normal until they were 12 or 15 or 18 months old, when an MMR shot -- or a combination of MMR and other shots -- led to their physical and mental collapse. Burton believes that a combination of 11 shots overwhelmed the immune system of his grandson Christopher, causing him to become autistic.

But the coincidence of MMR and autistic symptoms could be just that -- a coincidence. Ever since autism was described and named for the first time in 1944, clinicians have reported a large percentage of cases that developed during the second year of a child's life -- which also now happens to be the year of the MMR shots.

Public health authorities settled on the second year for MMR because they want young children, at the greatest risk for serious harm from measles, to be protected as soon as possible. The shot can't be given until a child is 1 year old because antibodies passed to the infant by its mother interfere in the workings of the shot.

When one talks to pediatricians and psychiatrists with years of experience treating autistic kids, the skepticism about the vaccine as cause is palpable, though few are willing to strike the notion altogether.

Sign up to receive free e-mail updates from Salon -- now in 17 different varieties!