Editor's Note: This article will appear in the July issue of Scientific American. We are posting it early in light of a report today in the journal Science that bears on similar themes. The study, led by Nancy J. Cox of the Centers of Disease Control and Prevention describes a molecular analysis of the novel influenza A (H1N1) virus infecting humans in several parts of the world. The authors confirm that the new strain is comprised of segments from swine flu strains known to circulate in Europe, Asia and North America, but that this combination has not previously been seen and appears to have been evolving independently from its parent strains for some time. Noting the "relative lack of surveillance for swine influenza viruses," the authors suggest, "this virus might have been circulating undetected among swine herds somewhere in the world." The study also confirms that the "H1" hemagluttinin protein of the new virus derives from the classical swine H1N1 strain, which shares a close common ancestor with the human H1N1 strain circulating before 1957 and several lines of evidence show that older people exposed to that virus may have some immunity to the new H1N1. Finally, the study finds that genetic changes associated with adaptation to a new host seen in other flu viruses that have breached the species barrier—such as the avian H5N1 flu strain—are not seen in the novel H1N1, suggesting that unknown changes in the new virus account for its ability to replicate and transmit in humans.

Less than 24 hours after a commercial jet took a sudden detour into the Hudson River this past January, security camera video of the event from multiple vantage points began surfacing. In an age of ubiquitous surveillance, the public has come to assume that someone or something is always watching, ready to spot trouble as it is happening. Yet a novel strain of the influenza A (H1N1) virus jumped species and burst into the human population in March and April, and by late May health and agriculture officials were still trying to figure out where it came from.

The emergence of the new H1N1 flu strain has demonstrated the effectiveness of existing systems to watch for human flu outbreaks while also proving a long-standing theory that pigs could serve as mixing vessels for a pandemic virus. But it has also highlighted how disappointing progress has been in detecting where and how such viruses evolve in animals and in predicting their transmission to people—abilities that might have helped avert a pandemic or at least provide an early warning.

Despite years of attention and funding for flu research, however, health officials are no closer to having an efficient way to flag new animal pathogens that could harm people. For example, in 2007, when Jürgen A. Richt and his colleagues at the U.S. Department of Agriculture’s National Animal Disease Center in Ames, Iowa, identified a new influenza A (H2N3) strain in pigs that they thought had pandemic potential, “there was no one to tell,” he recalls. “So we asked ourselves, ‘What do we do with it?’ Nobody was interested—there was no rule or regulation in place.” Richt and a group of collaborators published their assessment of the new strain in a scientific journal article that concluded that “it would be prudent to establish vigilant surveillance in pigs and in workers who have occupational exposure.”

In the context of disease, surveillance means at a minimum that doctors and diagnostic laboratories report every instance of certain pathogens they detect. All human flu cases are “reportable” to the Centers for Disease Control and Prevention, for example, which tracks the incidence and movement of the illness. But for both people and animals, voluntary lab testing to diagnose flu captures only the small fraction of cases that ever involve a doctor visit. Systematic sampling and mandatory reporting of disease in swine herds are limited to a handful of commercially devastating illnesses, including classical swine fever and nipah virus.

Richt, now at Kansas State University, thinks veterinary diagnostic labs could play an important role in more active animal screening by testing every sample submitted for any reason for a full spectrum of pathogens. “We need a better network to look in animal populations for emerging infectious agents, with 21st-century technology,” he says. Big state laboratories, including the ones in Iowa and North Carolina, home to the largest U.S. swine populations, already have the technical ability to screen for a range of pig diseases, Richt explains. Microarray chips able to test for pathogens specific to pigs, cattle or poultry could give smaller labs the same capacity and provide a more comprehensive, real-time picture of microbial threats to people, such as a new flu variant, brewing in livestock.

Identifying novel flu strains in animals is one thing; determining whether they pose a human danger is another. “I’m a lot more pessimistic about being able to predict these things,” says Jeffery K. Taubenberger of the National Institute of Allergy and Infectious Diseases. In March he published an analysis of two swine branches of the H1N1 family tree—one of them a Eurasian strain that contributed segments of the novel H1N1 now infecting humans. The two strains had a common H1N1-type ancestor, but both have been evolving independently in pig populations, and the minute changes to viral genes that allowed the virus to adapt to a new host were different in each strain. Many other scientists looking for consistent signals that a virus is changing hosts or is becoming more transmissible or more virulent have also failed to find clear patterns.

As a result, no one can explain why the avian H5N1 flu virus has infected some 400 people worldwide, mainly in Asia and Africa, but failed so far to adapt completely to humans. Nor do scientists know where the original 1918 pandemic virus came from or where its distant descendant, the new H1N1 strain, is going. Having spread to 40 countries, infected nearly 10,000 people and killed 79 as of late May, it might still fizzle in the coming months or learn to transmit between people more easily. And this fall it could return to the Northern Hemisphere as a lion or a lamb.

Taubenberger, who with his colleagues at the Armed Forces Institute of Pathology first fished the 1918 H1N1 pandemic strain out of preserved samples of victims’ tissue in 1996, says too much is still unknown about the basic biology and ecology of flu viruses. He thinks surveillance of an entire rural ecosystem—pigs, birds, people, as well as dogs, cats, horses, and other domesticated and wild animals—would finally yield some deeper insights into why and how flu viruses evolve.

Fortunately, money and research directed toward pandemic preparations have dramatically improved human flu surveillance and response systems. Richt points to how rapidly labs identified the first U.S. cases of the new flu in two children in southern California and alerted the CDC, allowing health officials to swing into action. Unfortunately, without closer monitoring of the animal sources of novel flu strains, human surveillance will have to remain the first line of defense.