A zoologist traces flu across the globe.

In winter, everybody recognizes a stuffy nose, a fever and an achy body as influenza. But experts still grapple with where the flu virus goes during the summer. One theory has it that flu lays low, holding out until the following season in a small number of asymptomatic people. Another idea — that flu strains tend to become extinct locally but shift around geographically — carries more weight. A recent paper by Derek Smith of the University of Cambridge, UK, and his colleagues helped nail the latter hypothesis by plotting the results of antigen-binding assays and genetic sequencing of more than ten thousand viruses on a map (C. A. Russell et al. Science 320, 340–346; 2008).

The researchers call this approach 'antigenic cartography'. Their antigenic time charts contain data crunched from the portion of the World Health Organization's enormous 'Global Influenza Surveillance Network' database that details strains classified as 'H3N2' between 2002 and 2007. First, they confirm flu's source–sink dynamics by showing that winter flu strains are more closely related to (and thus more likely to have evolved from) strains found elsewhere than to last season's local contagion. Second, the team pinned down H3N2's spread. Temperate regions are regularly seeded by strains from east and southeast Asia, where many strains circulate continuously and asynchronously in a pattern probably driven by varying climatic conditions.

These findings suggest that close surveillance of emerging strains in east and southeast Asia could enable us to predict those that will later affect the rest of the world. Yet it also poses a question: why do flu strains not return to this region after spending time (and thus evolving) elsewhere? Now that we know where new strains come from, we need to find out why they never go back.

Discuss this paper at http://blogs.nature.com/nature/journalclub