female purple martin
© Timothy J MortonA female purple martin wearing a miniaturized geolocator backpack and leg bands. The bird was one of 34 songbirds outfitted with the devices so that York University researchers could track their fall and spring migration for the first time. A colour-coded band helps researchers identify the bird on her return.

A novel way of fixing tiny recorders to songbirds has revealed how they dawdle on their migration south to their wintering grounds, then race northwards again in the spring. And birds that breed together may winter together too - a finding that may have profound implications for the conservation of dwindling songbird species.

Biologists have long wondered about the details of migratory songbirds' intercontinental wanderings, but tracking the tiny birds has proven nearly impossible.

Now Bridget Stutchbury, an ornithologist at York University in Toronto, Canada, and colleagues have found a way to do it.

In Pennsylvania during the summer of 2007, the researchers fitted 14 wood thrushes and 20 purple martins with fingernail-sized "backpacks" that record the time of sunrise and sunset each day.

Hurried return

When the birds returned the next spring from their winter migration to Central and South America, Stutchbury was able to recapture five thrushes and two martins, recover the devices, and calculate each bird's latitude and longitude each day. "This is the first time anyone in the world has been able to map songbird migration routes to the tropics and back," she says.

The birds often took lengthy stops during their southward flight, but they hastened directly back northward, so that the spring migration ended up two to six times faster than the autumn one, the team found. This is probably because birds returning early to their northern breeding grounds can claim the best territories and get the breeding off to an early start, says Stutchbury.

The migration records also showed that the five wood thrushes all wintered within a small area on the coast of Honduras, rather than scattering throughout the tropics. This suggests that each breeding population might have its own distinct wintering area.

If so, the loss of a single stretch of tropical forest could snuff out a breeding population completely, rather than reducing numbers slightly throughout the breeding range - a crucial difference in designing conservation programmes, says Jeff Kelly, a migration ecologist at the University of Oklahoma in Norman.

Journal reference: Science (DOI: [link], in press)