Two frequent assumptions about the evolution of long-distance migration in birds are that they travel long distances annually to reach food-rich areas for breeding, and that they time their migratory journey to be at staging sites when the latter provide the best feeding conditions. These assumptions have rarely been properly tested, and there is no study in which a species’ major food types have been measured by standardized methods throughout a flyway and over a large part of the year. We here present such data for Eurasian teal (Anas crecca), converted to a common energetic currency, and collected at wintering, spring staging and breeding sites. Teal did not time migration to maximize local food abundance; most birds left wintering and spring staging sites before a sharp increase in invertebrate food abundance occurred. On the other hand, hatching of ducklings coincided with a peak in invertebrate food abundance on boreal breeding lakes. Mean overall food abundance (invertebrates and seeds combined) did not differ between wintering sites in southern France and breeding sites in northern Sweden at the time of breeding. Our results are inconsistent with the hypothesis that long-distance migration in dabbling ducks has evolved because adult birds gain an immediate pay-off in increased food abundance by flying north in spring. However, our data confirm a selective advantage for breeding at higher latitudes, because hatching of ducklings may coincide with a peak in invertebrate emergence and because longer days may increase the duration of efficient foraging.