Migrating birds, flying nonstop over long distances, are substantially
heavier at the start than at the end of their journey. Aerodynamic mo
dels predict that these birds would optimally have to fly faster in th
e beginning of their flight, and end at a slower speed. Energy expendi
ture would be extremely high in the beginning, decreasing towards the
end. Trained kestrels fly slower when carrying a load, generating the
required extra Lift by changing the wingbeat kinematics. An allometric
equation, describing the relationship between empirically derived fli
ght costs at the maximum range speed and body mass, is used to calcula
te the flight range of a wader that loses more than 60% of its lean bo
dy weight during migration. Flight speed predictions are based on the
kestrel data. Results of this novel approach are shown to provide more
realistic predictions than those based on an aerodynamic model.