The use of celestial or geomagnetic orientation cues can lead migratory birds along different migration routes during the migratory journeys, e.g. great circle routes (approximate), geographic or magnetic loxodromes. Orientation cage experiments have indicated that migrating birds are capable of detecting magnetic compass information at high northern latitudes even at very steep angles of inclination. However, starting a migratory journey at high latitudes and following a constant magnetic course often leads towards the North Magnetic Pole, which means that the usefulness of magnetic compass orientation at high latitudes may be questioned. Here, we compare possible long-distance migration routes of three species of passerine migrants breeding at high northern latitudes. The initial directions were based on orientation cage experiments performed under clear skies and simulated overcast and from release experiments under natural overcast skies. For each species we simulated possible migration routes (geographic loxodrome, magnetic loxodrome and sun compass route) by extrapolating from the initial directions and assessing a fixed orientation according to different compass mechanisms in order to investigate what orientation cues the birds most likely use when migrating southward in autumn. Our calculations show that none of the compass mechanisms (assuming fixed orientation) can explain the migration routes followed by night-migrating birds from their high Nearctic breeding areas to the wintering sites further south. This demonstrates that orientation along the migratory routes of arctic birds (and possibly other birds as well) must be a complex process, involving different orientation mechanisms as well as changing compass courses. We propose that birds use a combination of several compass mechanisms during a migratory journey with each of them being of a greater or smaller importance in different parts of the journey, depending on environmental conditions. We discuss reasons why birds developed the capability to use magnetic compass information at high northern latitudes even though following these magnetic courses for any longer distance will lead them along totally wrong routes. Frequent changes and recalibrations of the magnetic compass direction during the migratory journey are suggested as a possible solution.