Large migratory birds may bring along stores in order to survive adverse conditions and produce a clutch upon arrival (‘capital breeders’) or they may acquire all the necessary resources on the breeding grounds (‘income breeders’). Whether birds are capital- or income-breeders may depend on the distance between the last stopover site and the breeding grounds and the length of the summer season. The degree of capital-breeding may therefore differ among flyway populations of the same species. I used migration speed as a proxy for the degree of capital-breeding, and compared the observed migration speeds of satellite-tracked Tundra Swans Cygnus columbianus from all four main flyway populations with those predicted by an allometric model for 6-kg waterfowl. Average overall spring migration speed of Tundra Swans was 52.2 km/d, close to that predicted under a capital-breeding strategy (i.e. carrying the stores for the whole clutch and allowing for 8 days of starvation). This migration speed was in accordance with the speed at which the ice retreats in spring. However, the energy stores for clutch formation and survival on the breeding grounds may be primarily put on at the last stopover. Tundra Swans (‘Whistling Swans’) in the Nearctic, and especially in the Western Nearctic, make use of the rapid advancement of spring on their last leg and accelerate their migration speed accordingly, and so probably rely on income-breeding. In the Western Palearctic, Tundra Swans (‘Bewick’s Swans’) are faced with a slow advancement of spring on their last track, and they may fare better by building up stores on the last stopover and bring these to the breeding grounds. This may not be an option for swans in the Eastern Palearctic, as they are confronted with a very long last leg. The variation in migration speeds on the last leg suggests a large variation in the degree of capital-breeding among flyway populations of Tundra Swans.