Since 1909, researchers have been catching and marking migrating birds that stop over on the island of Helgoland in the southeastern North Sea. These birds breed in Scandinavia and spend the winter in either continental Europe (short-distance migrants) or Africa (long-distance migrants). The methods of trapping — one type of apparatus is shown in the photograph opposite — have not changed since 1960. Moreover, the data cover around two dozen species, and describe the mean time of migration for all trapped individuals, not just extremes in the form of first arrivals. All of this makes the Helgoland data sets some of the best available with which to study the timing of bird migration.

Ommo and Kathrin Hüppop have now analysed these data sets, with remarkable results (Proc. R. Soc. Lond. B 270, 233–240; 2003). They find that all 23 migratory bird species for which sufficient data are available pass by Helgoland on their way to Scandinavia earlier now — by two to twelve days — than 40 years ago. There is a clear division between short-distance migrants, whose mean time of passing correlates well with local temperatures, and long-distance migrants, for which increases in the NAO index (a measure of the air pressure over the North Atlantic Ocean) give a much better explanation for the earlier time of passage.

Whether a migrating bird actually lands on a small island while passing over it depends on many factors. Sudden changes in weather play a large part. So it is important to analyse large data sets to disentangle general patterns from such isolated examples. The changes in the timing of migration are apparently strong enough to become evident.

These changes, in particular the earlier passage of the long-distance migrants, raise questions about the control of spring migration. There are three main hypotheses that might explain their earlier passage. First, the moment of leaving Africa has not changed, but refuelling in continental Europe proceeds more quickly, because more food — in the form of insects — is available earlier. (Increases in the NAO index generally indicate favourable spring conditions in Europe.) Second, if the weather in Africa is also correlated with the NAO index, then the birds might leave earlier because the seasons there also change earlier. Third, the weather in Africa has not changed, but natural selection has altered the 'trigger values' for starting migration. Each hypothesis, if true, would mark an exciting break with existing knowledge, and I eagerly await further results on changes in spring migration.