A global influenza pandemic is closer than at any time in a generation. Klaus St¿hr, head of flu surveillance at the World Health Organization, made that uncharacteristically dramatic declaration back in November, before convening an emergency summit of vaccine manufacturers and national health agency officials. The reason was the escalating avian influenza crisis in 10 Asian countries, which poses a direct threat to human health.

Because the avian virus, known as H5N1, is lethal in chickens, identifying local outbreaks was easy. St¿hr was alarmed by evidence that the virus is also widespread in the region's domestic ducks, which show no symptoms. With asymptomatic ducks waddling through barnyards, backyards and kitchens shedding virus, hope of stamping out this largest avian flu outbreak in history has dimmed. And the likelihood of human exposure to the virus, which could spark a pandemic, has increased. "We know the recipe, and all the ingredients are there," St¿hr said.

Flu pandemics are caused by viruses with surface proteins unfamiliar to human immune systems. In the pandemic strains of 1957 and 1968, such proteins came from flu viruses that previously infected only birds. That was possible because flu is a promiscuous gene swapper. If an avian strain meets a flu strain adapted to spreading among mammals, their progeny may possess the deadly combination of protein novelty and easy transmissibility between humans. The exchange can happen inside animal hosts susceptible to both avian and human flu strains, such as pigs. Or it might happen inside a farmer's child infected with this season's circulating human flu strain, whose healthy-looking pet duckling then gives her H5N1. In the past year more than 40 people in Thailand and Vietnam have already contracted the avian virus, and more than 30 of them have died from it.

Beginning on page 62 of this issue, Jeffery K. Taubenberger and his colleagues describe their resurrection, literally from the grave, of the virus that killed upward of 40 million people in 1918¿1919. And they offer a chilling conclusion about its origin. Although certain genes in that virus may have come from a bird strain, the genes look as though they spent significant time evolving in yet another animal before the virus emerged among humans. This unknown source of the 1918 pandemic might even have been a type of bird or animal that we don't recognize as a flu carrier.

Systematic surveillance for influenza is currently limited to humans, chickens, swine and horses, with wild waterfowl and shore birds tested less regularly. But few of these samples are closely analyzed, and scientists do not know the dynamics or true scope of the influenza ecosystem. The National Institutes of Health just announced plans to sequence the genomes of existing flu samples. In addition, the Armed Forces Institute of Pathology, the Centers for Disease Control and Prevention, and several partners want to create a high-capacity laboratory network dedicated to expanding flu surveillance capability and molecular analysis of influenza viruses. Studying how they evolve and move through human and animal populations might identify new flu reservoirs and enable the emergence of new strains to be predicted.

Start-up costs would equal the price of the two million emergency H5N1 vaccine doses the Department of Health and Human Services ordered in September. The money would be well spent for an ounce of prevention against future flu pandemics.